Search…
Summarize
Summarize any length of text or dialogue.

Create summary

Forefront Summarize is an abstractive text summarization API that generates high-quality text or dialogue summaries. See pricing
post
https://solutions.forefront.ai
/v1/organization/{TEAM_ID}/summarize
Create summary

Example request

Python
Javascript
cURL
import requests
import json
headers = {
"Authorization": "Bearer {API_KEY}",
"Content-Type": "application/json"
}
body = {
"compression_level": 3,
# Send text or dialogue to summarize, but not both
"text": "<Enter text to summarize>",
"dialogue": [
{
"speaker": "<enter name>"
"utterance": "<enter utterance>"
},
{
"speaker": "<enter name>"
"utterance": "<enter utterance>"
},
...
]
}
res = requests.post(
"https://solutions.forefront.ai/v1/organization/{TEAM_ID}/summarize",
json=body,
headers=headers
)
data = res.json()
const fetch = require('node-fetch');
const url = 'https://solutions.forefront.ai/v1/organization/{TEAM_ID}/summarize';
const headers = {
"Authorization": "Bearer {API_KEY}",
"Content-Type": "application/json"
};
const body = {
"compression_level": 3,
// Send text or dialogue to summarize, but not both
"text": "<Enter text to summarize>",
"dialogue": [
{
"speaker": "<enter name>"
"utterance": "<enter utterance>"
},
{
"speaker": "<enter name>"
"utterance": "<enter utterance>"
},
...
]
}
fetch(url, {
method: 'POST',
headers,
body: JSON.stringify(body)
}).then(async (res) => {
const data = await res.json()
console.log(data)
})
curl https://solutions.forefront.ai/v1/organization/{TEAM_ID}/summarize \
-H 'Content-Type: application/json' \
-H 'Authorization: Bearer {API_KEY}' \
-d '{
"compression_level": 3,
"text": "<Enter text to summarize>",
"dialogue": [
{
"speaker": "<enter name>"
"utterance": "<enter utterance>"
},
{
"speaker": "<enter name>"
"utterance": "<enter utterance>"
},
...
]
}'

Example use cases

The Forefront Summarize API can be used to accurately summarize text or dialogue at different lengths.

Summarize dialogue

An example use case for summarizing dialogue would be to summarize an earnings call transcript. Take the following Q3 2021 Ford Motor Company earnings call:
Ford Motor Company Q3 2021 earnings call (56,615 characters)
First, we will convert the transcript to the format required by the API: an array of objects with speaker and utterance key-value pairs. Once formatted properly, we can summarize all of the dialogue in the following API call:
import requests
import json
headers = {
"Authorization": "Bearer {API_KEY}",
"Content-Type": "application/json"
}
body = {
"compression_level": 5,
"dialogue": [{"speaker": "Erica", "utterance": "Good day, ladies and gentlemen. My name is Erica, and I will be your conference operator today. At this time, I would like to welcome you to the Ford Motor Company third-quarter 2021 earnings conference call. All lines have been placed on mute to prevent any background noise."}', '{"speaker": "Erica", "utterance": "After the speakers\\u2019 remarks, there will be a question-and-answer session. If you would like to ask a question during this time, you would need to press *1 on your telephone."}', '{"speaker": "Erica", "utterance": "At this time, I would like to turn the call over to Lynn Antipas Tyson, Executive Director of Investor Relations."}', '{"speaker": "Lynn Antipas Tyson", "utterance": "Thank you, Erica, and welcome to Ford Motor Company\\u2019s third-quarter 2021 earnings call. With me today are Jim Farley, our President and CEO; and John Lawler, our Chief Financial Officer. Also joining us for Q&A is Marion Harris, CEO of Ford Credit."}', '{"speaker": "Lynn Antipas Tyson", "utterance": "Today\\u2019s discussions include some non-GAAP references. These are reconciled to the most comparable U.S. GAAP measures in the appendix of our earnings deck. You can find the deck along with the rest of our earnings materials and other important content at shareholder.ford.com, including some updated videos and proof points around our Ford+ plan for growth."}', '{"speaker": "Lynn Antipas Tyson", "utterance": "Today\\u2019s discussion also includes forward-looking statements about our expectations. Actual results may differ from those stated. The most significant factors that could cause actual results to differ are included on page 23. Unless otherwise noted, all comparisons are year-over-year. Company EBIT, EPS and free cash flow are on an adjusted basis, product mix is volume weighted."}', '{"speaker": "Lynn Antipas Tyson", "utterance": "A quick update on our IR events for the balance of the year, we have five. On Monday the 1st, November 1st, Wolfe will host a fireside chat with John Lawler and Hau Thai-Tang, our Chief Product Platform and Operations Officer. On the 18th of November, Barclays will host a virtual fireside chat with Ted Cannis, our CEO of Ford Pro. In December, on the 3rd, Goldman Sachs will host a virtual fireside chat with Lisa Drake, our Chief Operating Officer for North America. On the 3rd, Credit Suisse will host fireside chat with Hau Thai-Tang. And finally, on the 9th, Deutsche Bank will host a virtual fireside chat with Alex Purdy, our Director of Business Operations Enterprise Connectivity."}', '{"speaker": "Lynn Antipas Tyson", "utterance": "Now I\\u2019ll turn the call over to Jim Farley."}', '{"speaker": "James Farley", "utterance": "Thank you, Lynn. Hello, everyone, and thanks for joining us today. While, this month marks one year since we began executing our Ford+ plan. It creates growth, value and it allows us to win in the emerging area of electric and connected vehicles. We built a strong team that combines top leaders from Ford with world-class talent recruited from outside of our company, specific talent, specific people that are largely outside of our industry. This leadership team is committed to this plan, and we are accelerating our progress."}', '{"speaker": "James Farley", "utterance": "Ford+ is not a tag line. It\\u2019s not advertising. It\\u2019s a larger, more ambitious way to think about our business and how we bring value to our customers. We\\u2019re creating iconic and distinctive products that only Ford can do; increasingly, always on relationships with our customers and ever improving user experiences, all while creating value for our shareholders. We\\u2019re fully vested in this future and we\\u2019re taking big swings. Key to the plan is harnessing the power of connectivity. We\\u2019re designing a new generation of fully networked vehicles not delegated to our supply chain but done inside the company to revolutionize the experience of owning and operating Ford vehicle. That\\u2019s embedded technology to unleash unlimited innovation."}', '{"speaker": "James Farley", "utterance": "We\\u2019re scaling the number of vehicles capable of over their updates. We\\u2019ll be moving from 1 million vehicles today to 33 million by 2028. That\\u2019s scale. At the same time, we\\u2019re moving aggressively to lead the electric vehicle revolution, substantially expanding our battery production as we speak today in US. In fact, we already announced plans that will give us enough battery production to meet our mid-decade goal of 141 gigawatts, which is enough to build more than 1 million battery electric vehicles a year, and I think we\\u2019ll need more."}', '{"speaker": "James Farley", "utterance": "Our $7 billion investment in BlueOval City in Tennessee and the BlueOval SK battery park in Kentucky sets a new standard for scale, sustainability, advanced manufacturing and training the next generation of technology leaders. At the same time, $1 billion of investment in our Electrified Center in Cologne, Germany, will allow us to go all-electric soon. That center will be all electric by 2023."}', '{"speaker": "James Farley", "utterance": "We\\u2019re making final preparations to launch the F-150 Lightning, the defining zero emission version of America\\u2019s best-selling vehicle for the past 40 plus years. Our all-electric Mustang Mach-E is a hit with customers not just in the US, but around the world, bringing a stunning number of new customers to the Ford brand. Over 90% of the Mach-E owners say they would recommend a Mach-E to other customers, critical as a new generation of battery-electric customers make new brand choices. Our challenge now is to break production constraints and increase availability to meet this incredible demand both in North America and in Europe and also in China, the biggest EV market in the world where we are just starting production of Mach-E. We believe the global demand just for Mustang Mach-E could approach about 200,000 vehicles a year."}', '{"speaker": "James Farley", "utterance": "We\\u2019ve created new organization, Ford Pro, to change and power the future of work with compelling commercial vehicles, distribution and services, while growing revenues to Ford. In the few weeks, we will start production of the new E-Transit, an electric version of the world\\u2019s best-selling commercial van. And in the third quarter alone, FORDLiive, Europe\\u2019s new connected uptime center, which I wish you could all see, for our commercial customers, helps customers in the UK secure additional uptime, preventing about $8 million of lost revenue and associated cost for our valuable customers."}', '{"speaker": "James Farley", "utterance": "We\\u2019re reinventing icons like Bronco and creating new ones like Maverick. In fact, the all-new Maverick, 42 miles per gallon, I might add, is the first standard hybrid pickup in the United States. It\\u2019s also America\\u2019s most fuel-efficient hybrid pickup. This is the strongest, most compelling lineup I\\u2019ve ever seen from any mass-market brand in my career. And we are creating a spring-loaded future as we emerge from the chip shortages and COVID constraints."}', '{"speaker": "James Farley", "utterance": "We continue to make important strides in the technology and go-to-market strategy for autonomous vehicles. In the second quarter, we told you about a new partnership with Argo AI and Lyft. And in the third quarter, we announced a new partnership with Argo AI and Walmart. This is Walmart\\u2019s first-ever multi-city autonomous delivery service, and it will be anchored in cities where we already have operations; not the easiest miles in one city, but multiple cities and hard miles. In addition to making real progress on autonomy vehicles, operating domains, SDS, we fully support Argo AI\\u2019s aspiration to access public capital."}', '{"speaker": "James Farley", "utterance": "To build this future and generate the margins and cash flow we need to fund Ford+, we had to turn around our automotive operations and improve our competitiveness. Our results in the third quarter show we are making significant progress. In fact, company-wide, we achieved an 8.4% EBIT margin, including 10.1% in North America. Those margins, I will remind you, are in line with our targets for 2023. More importantly, our operations outside of North America are likely to post the best performance in four years. Please note the performance in South America, largely driven by our success of our global redesign. We\\u2019ve been able to achieve this while thoughtfully managing our supply chain for short-term sustainable improvements, including semiconductors and prioritizing high demand and high profit vehicles."}', '{"speaker": "James Farley", "utterance": "Now before I turn it over to John, a few thoughts. I believe we have the right plan to drive growth and unlock unprecedented value. You are already seeing the favorable change in the slope of our earnings and cash flow. There is more to come. Given the strength of our business this year, we are increasing our full year adjusted EBIT guidance to between $10.5 billion and $11.5 billion."}', '{"speaker": "James Farley", "utterance": "As we plan in earnest for next year, we\\u2019re excited and energized about the opportunity in front of us and clear that we have so much more work to do to deliver on Ford\\u2019s potential. The word I would leave you with is focus. The competitive environment has never been more interesting and tough. And we intend to live up to our promise to compete like a challenger, focusing on our top priorities to unlock Ford+ growth with customers at the very center of everything we do."}', '{"speaker": "James Farley", "utterance": "Now I\\u2019ll turn it over to John who will take you through our results for the quarter, our outlook for the full year, our capital allocation priorities, and our expectations heading into next year. John?"}', '{"speaker": "John Lawler", "utterance": "Thank you, Jim. Now in the face of continued industry-wide semiconductor constraints, we stayed focused on our plan, strengthening our portfolio, and investing in opportunities fundamental to growth and value creation. We delivered a solid quarter with $3 billion in adjusted company EBIT and a margin of 8.4%. Free cash flow of $7.7 billion was as we expected, up sharply on a sequential basis, driven by the positive working capital effects from higher wholesales in EBIT. We ended the quarter with strong cash and liquidity at over $31 billion and $47 billion respectively."}', '{"speaker": "John Lawler", "utterance": "Now across our automotive business, our playbook remained consistent as we optimized production for customer orders, new launches in our most profitable vehicles. And as expected, on a sequential basis, our wholesales improved dramatically as chip supply for Ford improved. We also remain disciplined with incentive spending and mix management, which on a year-over-year basis, more than offset chip-related declines in volume."}', '{"speaker": "John Lawler", "utterance": "Ford Credit delivered another solid quarter with $1.1 billion in EBT as auction values continued to remain strong and credit losses continue at near-record lows. For the fourth quarter, we\\u2019re assuming a sequential increase in wholesales. We also expect continued healthy mix and net pricing and solid results from Ford Credit, although not as strong as the third quarter. Headwinds include inflationary impacts on commodities and freight, and we also expect planned sequential increases in our Ford+ modernization investments, including customer experience and IT."}', '{"speaker": "John Lawler", "utterance": "Now let me share with you some highlights from the quarter before I turn to guidance, capital allocation, and our preliminary view of 2022. With improved chip supply, North America wholesales increased sequentially by 67% as the team prioritized launches, customer orders, and high margin units, while reducing the number of vehicles built, but waiting for chips. Demand remained strong for our exciting vehicles. The order bank we are building paid off in the third quarter, representing 28% of our retail sales in the third quarter and reaching a high of 31% in September. And our overall customer orders increased over 50% from the second quarter to more than 100,000 orders, excluding Bronco. With a 10.1% EBIT margin in the quarter, North America is now at a 9% margin year-to-date, just 100 basis points shy of our 2023 target of 10%."}', '{"speaker": "John Lawler", "utterance": "South America marked its eighth consecutive quarter of year-over-year improvement in EBIT, and the business run rate is now approaching breakeven. The region also launched its new commercial vehicle organization with the introduction of the new transit, which is manufactured in Uruguay. This transit is the first light commercial van to market in Brazil that includes connectivity as a standard feature."}', '{"speaker": "John Lawler", "utterance": "In Europe, the underlying trajectory of our business continues to strengthen, though the adverse impact from chips has masked this improvement. In the third quarter, the business lost about 50,000 units, which would have had a substantial favorable impact on EBIT. Our leadership, as the number one commercial vehicle brand, continued in the quarter along with an extremely robust order bank."}', '{"speaker": "John Lawler", "utterance": "In China, Lincoln continues to perform well, extending its success in the most profitable segment, Luxury. With retail sales up 24% year-over-year, in fact, Lincoln has doubled its share of the China luxury market over the past 18 months. Our newly created BEV organization opened its first 13 direct-to-consumer Ford Select city stores with a total of 25 expected to open by year end."}', '{"speaker": "John Lawler", "utterance": "In IMG, our leadership team in India made the difficult decision to end manufacturing following accumulated operating losses of more than $2 billion over the past 10 years. Going forward, we will focus on importing iconic vehicles including EVs, and overall, IMG had a solid quarter, capitalizing on our strengths, including Ranger."}', '{"speaker": "John Lawler", "utterance": "Now as Jim highlighted, the underlying strength of our business supports increasing our adjusted EBIT guidance for 2021 to between $10.5 billion and $11.5 billion, and that\\u2019s despite a lower than anticipated improvement in chip availability in the second half of the year. Consistent with our adjusted EBIT guidance through this year, our updated guidance for 2021 includes the $900 million non-cash gain on our investment in Rivian in our first quarter adjusted results."}', '{"speaker": "John Lawler", "utterance": "Let me spend a minute on Rivian. Now in the event that Rivian completes its IPO, we will record any gain on our investment in any subsequent adjustments as special items. Accordingly, we will recast the $900 million non-cash gain from adjusted EBIT in the first quarter to a special item. If Rivian completes their IPO in the fourth quarter, we will make this change when we report our fourth quarter earnings on February 3, 2022."}', '{"speaker": "John Lawler", "utterance": "Our guidance for 2021 adjusted free cash flow is unchanged at $4 billion to $5 billion, reflecting the higher EBIT, but less favorable improvement in working capital and timing differences. Now this is due to lower than anticipated volume than previously assumed in the back half of the quarter, and that\\u2019s as a result of chip constraints. We do expect free cash flow to increase with higher production and the associated improvement in supplier payables and other timing differences."}', '{"speaker": "John Lawler", "utterance": "Now let me turn to capital allocation, which again, is the foundation of our value creation framework. Our capital allocation discipline is driving a strong core business and balance sheet that provides the flexibility to invest in new growth opportunities as we deliver our Ford+ plan. Ultimately, it ensures we return value to our shareholders, both in the form of a higher share price and dividends."}', '{"speaker": "John Lawler", "utterance": "Today, we announced the reinstatement of our dividend. Our board has approved restarting a regular quarterly dividend of $0.25 per share in the fourth quarter of 2021. Importantly, the dividend reflects our confidence in the improving run rate of the business and our ability to fund all of our calls on capital, including the growing investment in electrification and the trajectory of our Ford+ plan. The dividend was also sized to ensure we maintain appropriate optionality to manage continued uncertainties in the external environment."}', '{"speaker": "John Lawler", "utterance": "To give you a better sense of our calls on capital between 2020 and 2025, we expect total capital expenditures of about $40 billion to $45 billion or a run rate of roughly $7 billion per year. Now over the same time, we expect to invest over $30 billion in BEVs. And of the investment in BEVs, about 50% is capex, 25% is expense, and 25% is direct investments. And these numbers, they will be dynamic and we are confident we have ample financial flexibility to increase our investments even if BEV adoption further accelerates."}', '{"speaker": "John Lawler", "utterance": "Now let me share with you our early thinking about 2022, a year which like this one, is likely to experience some industry crossroads that could drive a range of outcomes. Now we typically don\\u2019t talk about the upcoming year this soon, and were not yet prepared to give financial guidance, but we do want to share how we\\u2019re thinking about next year given the dynamic operating environment."}', '{"speaker": "John Lawler", "utterance": "Ford\\u2019s underlying strengths give me great confidence we can build on our results in 2021. First, our portfolio of products and services is exceptional, and we have a significant amount of new products coming to market spanning our iconic high volume nameplate. Second, our industrial base gives us significant optionality as the adoption of electric vehicles accelerates. Third, driven by the chip shortage of roughly 4 million in wholesales we are likely to deliver this year, fall significantly below our capacity. And based on our current assessment, we believe our wholesales to be up about 10% in 2022, but that number is very dynamic and changes almost weekly. And fourth, the effects of our global redesign, which is largely completed are now evident and substantial. We have drastically de-risked and rationalized our global footprint and product lineup, vastly improving our earnings and cash generation power in the process."}', '{"speaker": "John Lawler", "utterance": "Now for headwinds next year, it\\u2019s difficult to predict the interplay between semiconductor-related constraints, volume and pricing, and this will continue to remain dynamic. For 2021, we expect commodities to be up three to $3.5 billion and they could be up another $1.5 billion in 2022, largely driven by steel and aluminum similar to this year. There will also likely be other inflationary costs, but it\\u2019s too early to size that right now."}', '{"speaker": "John Lawler", "utterance": "Ford Credit is likely to be lower as strong auction values will be moderated by a smaller inventory of vehicles and lower lease and return rates. Lastly, we\\u2019re obviously going to continue to invest in our Ford+ plan for growth and value creation. And this includes, in customer-facing technology, connectivity, our always on relationships with customers and electrification. And of course, we believe the long-term payback from those investments will be substantial."}', '{"speaker": "John Lawler", "utterance": "Now that wraps up our prepared remarks. And if you perceive that the upfront portion of these calls is becoming more efficient, well, you\\u2019re right. That\\u2019s a function of us being very specific with you and our team about what\\u2019s truly important and our confidence in executing effectively against those things and reporting accordingly. We\\u2019ll use the balance of the time to hear and address what\\u2019s on your minds. Thank you."}', '{"speaker": "Erica", "utterance": "As a reminder, to ask a question, you would need to press *1 on your telephone. To withdraw your question, press the pound key. We do ask that you please limit yourself to one question and one follow-up question. Please stand by while we compile the Q&A roster."}', '{"speaker": "Erica", "utterance": "Your first question comes from the line of John Murphy with Bank of America."}', '{"speaker": "John Murphy", "utterance": "Good evening, everybody. Thanks for making the call the efficient. It\\u2019s going to be tough to limit to one question, but I will. As you think about the 10% increase you\\u2019re talking about in 2022 wholesales, and if we could focus on North America and just assume you\\u2019re going to do about 2 million units this year in \\u201921, give or take, we\\u2019re only talking about 200,000 units of increase next year. There is an assumption that price and mix will deteriorate as incremental units are produced as semi shortage is relieved. But given that that\\u2019s still going to be a relatively low\\u2026 a very low level of production, do you believe that price and mix are really going to actually come under pressure next year? And aren\\u2019t we really going to stay in a very tight environment that you\\u2019re selling through and not even building inventory, if that\\u2019s true, which means that price and mix might stay very strong next year and you will still get the benefit?"}', '{"speaker": "John Lawler", "utterance": "Thanks, John. John here. You\\u2019re right, it\\u2019s going to remain dynamic, and that\\u2019s what the interplay is going to be. Volume increases for the industry, if they\\u2019re higher, we\\u2019ll probably see more pressure on price. If they remain as they are today and we see a moderate increase, I think you\\u2019re going to continue to see strong pricing and mix continue through next year. So that\\u2019s where we have to stay disciplined and we have to stay very focused on managing that well so that we can have, as you said, the play through next year relative to what happens from an overall volume standpoint, and we\\u2019re focused on that. So I agree with you. That\\u2019s going to be one of the key dynamic elements for next year."}', '{"speaker": "John Murphy", "utterance": "And just to follow-up on that. I think right now based on awards, you have about 213,000 units in dealer inventory. Your pre-COVID travel rate was about 650,000 units. As you think about ultimately getting into a time where you can rebuild or restock that inventory, where do you think that runs? And how much opportunity is there to try to maintain some of this mix and price discipline to offset any cost inflation and then also invest in the future?"}', '{"speaker": "John Lawler", "utterance": "Yeah. If you look at where we ended in September, we ended at about 20 days supply in the US and we\\u2019re watching it very closely from a day supply standpoint. As we talked about last quarter, our historical day supply was somewhere around 75 days. We\\u2019re not going back there. As Jim said, we\\u2019re going to be very disciplined and we expect to be in the 50-day supply when we\\u2019re fully, fully at full capacity and we\\u2019re running and producing everything that we can. So that\\u2019s going to be the key."}', '{"speaker": "John Lawler", "utterance": "The other thing I would say, John, is that the move, as we talked about in our remarks, to having more of our sales come through orders, online orders in the order bank. That\\u2019s really important for us to manage our day supply. So it\\u2019s a less of a ground stock pushed through and it\\u2019s customer demand pull through based on the orders, and we had over 100,000 orders at the end of the quarter and that\\u2019s grown since then in our order bank and that played well for us in the third quarter."}', '{"speaker": "James Farley", "utterance": "139,454 orders-"}', '{"speaker": "John Lawler", "utterance": "As of-"}', '{"speaker": "James Farley", "utterance": "\\u2026 as of today."}', '{"speaker": "John Lawler", "utterance": "As of today."}', '{"speaker": "James Farley", "utterance": "To be exact. [crosstalk 00:25:30]."}', '{"speaker": "John Murphy", "utterance": "It sounds like you\\u2019re on that. And maybe if I could sneak one in, just on cap allocation, a real quick on the dividend, I mean, why now? And then I\\u2019ll hop back in the queue."}', '{"speaker": "John Lawler", "utterance": "Yeah. It\\u2019s just the underlying strength of the business, John, and we\\u2019re not capital constrained. We\\u2019re able to fund our initiatives for growth. We know that there aren\\u2019t going to be other opportunities that surface. We\\u2019re confident that we can fund those. And we\\u2019re focused on total shareholder returns. Not only stock appreciation, but also the dividend. And so given the strength of the business, the board elected that we would restart the dividend this quarter."}', '{"speaker": "John Murphy", "utterance": "Great. Thank you very much."}', '{"speaker": "Erica", "utterance": "Your next question comes from the line of Dan Levy with Credit Suisse."}', '{"speaker": "Dan Levy", "utterance": "Hey, good evening. Thank you for taking the question. I would just like to ask a question on the shape of recovery in volumes. A, is this magnesium shortage going to cause any sort of near-term supply disruption for you? I mean, we\\u2019ve heard some draconian comments that it could just outright, stop European production. Just maybe you can give us a sense broadly of how you anticipate the shape of recovery in terms of volumes? At what point do we get\\u2026 Or is your baseline expectation that the chip shortages fully mitigated that you can be back at full run rate production?"}', '{"speaker": "James Farley", "utterance": "Yeah. From the magnesium standpoint, when we look at that, we are seeing price pressure on aluminum broadly. We saw that all year and probably see a little bit more price pressure due to the magnesium issue. But our sheet metal suppliers, our sheet aluminum suppliers don\\u2019t purchase magnesium from China for North American production. So we don\\u2019t see that having any significant impact or any impact on us."}', '{"speaker": "James Farley", "utterance": "We do see the chip issue continuing to run through \\u201922. As we said, it\\u2019s very dynamic. Right now, you ask us what we think the sequential increase in supply will be year-over-year. We think we\\u2019ll have about 10% more, but that\\u2019s changing weekly, and we\\u2019re doing everything we can to get our hands on as many chips as we can. But we do see that running through 2022. It could extend into 2023, although we do anticipate the scope and severity of that to reduce as we move through \\u201922 into \\u201923."}', '{"speaker": "Dan Levy", "utterance": "Great. Thank you. The second question. I\\u2019d like to do now sort of a bit more strategic. I think if we just look at the pace of progress at Ford, just both financially and in terms of EV, AV, all things digital, is just what we\\u2019re seeing today is just it\\u2019s far greater than what we saw 12 or 18 months ago, which is actually pretty impressive given that large organizations just takes time to really affect change. So I\\u2019m trying to understand, how much of what we\\u2019re seeing today was something that was always there and it\\u2019s just only starting to come to the surface now, we\\u2019re sort of getting the fruits of prior initiatives? Or has something fundamentally changed in the past 12, 18 months? What does this tell us about the pace of change at Ford and as a comparison? Maybe just to be a bit more specific, because I know you can take that in number of ways, maybe you could answer that specifically to what we\\u2019re seeing on product and planning on the business transition to EV?"}', '{"speaker": "James Farley", "utterance": "Thank you for your question. A lot of the product we\\u2019ve been working on for several years, we made the tough choices. I would say the answer is, we have a plan. It\\u2019s not advertising or PR tag line, it\\u2019s our plan. Everyone in the company knows what we have to do. We are out of time and we have focus. We need to get 8% margin, like we did this quarter as a company regularly because we have to fund the high growth BEVs and digital business. It\\u2019s not to make more money. Yes, it is that, but it\\u2019s motivated in the mission of transforming Ford through these digital products."}', '{"speaker": "James Farley", "utterance": "So running the ICE business for cash, getting serious about our cost, our quality, our launches, our 8% return, it\\u2019s all mission and the team knows the plan. I think the culture is starting to change, to be quicker, more accountability, less bureaucracy and that mission permeates through the company. I\\u2019m probably the worst person to ask whether something has changed because it sure has changed for me and my leadership team. But to me, the proof is in the pudding, like our third quarter. And whether we\\u2019ve really changed as a company will be proven out in our numbers over time like they have the last year."}', '{"speaker": "Dan Levy", "utterance": "And just to be more specific on the product front, because I think we\\u2019re seeing a much faster pace of product. Is the time of development products\\u2026 How much have you accelerated that? Meaning, typically in the past we would hear of three and a half to four years of drawing board to product in showroom. Is there a new normal for what that is?"}', '{"speaker": "James Farley", "utterance": "If we make up our mind and we come together as a team, like we did on Maverick, it could be just two years. We knocked 20 months after the Maverick development. They required the leadership team to not have the hand wringing on the studio for six months like we normally did. I think that\\u2019s a new proof point. But the question I ask myself is a little different. When we see a technology change like this, like BEV, it\\u2019s not just the speed of your product creation, it\\u2019s can you be flexible and agile in your industrial system like in manufacturing?"}', '{"speaker": "James Farley", "utterance": "We have three complete hits on our hands; a Mach-E with 200,000 units of demand, we have the Lightening with over 160,000 orders, and the E-Transit is completely sold out. So how I like to think about it is not just the product creation speeding up, it\\u2019s whole company. And we have to do our job to break constraints now so that we can deliver hundreds of thousands of battery-electrics next year. That to me is the proof of our change, not just how fast the product creation process works."}', '{"speaker": "Dan Levy", "utterance": "Great. Thank you very much. Very helpful."}', '{"speaker": "Erica", "utterance": "Your next question comes from the line of Ryan Brinkman with J.P. Morgan."}', '{"speaker": "Ryan Brinkman", "utterance": "Hi. Thanks for taking my question. I thought to ask a few on the order bank, just given the commentary that it grew 50% sequentially in 3Q, excluding the Bronco. So can you talk about the benefits of the order bank? How it helps to optimize your operations? And what kind of pricing or other trends you might be seeing with regard to the order bank? And then how much of the increase in orders do you think may stem from the currently very strong new product cadence or from the currently low inventory environment? And what avenues are there available to drive orders as industry conditions eventually normalize?"}', '{"speaker": "James Farley", "utterance": "Thank you so much for your question. I\\u2019ll ask John to comment. From my view, the order bank model that we\\u2019re going to in North America that we\\u2019re in right now has benefits across the patch. We are an incredibly complicated company. And so having an order bank allows us to push simplification into the order, the customer-facing options, which we need to do, and it reduces cost and improves our quality."}', '{"speaker": "James Farley", "utterance": "Number two, it eliminates the need for expensive conquest-fixed marketing. Number three, it\\u2019s incredibly helpful for industrial system. You cannot imagine, Ryan, how much money we waste by guessing what our launch mix is for a new product. When you have a order bank, especially for new models, you could capacitize the high-series mix that are very profitable right in line with customer demand. So it\\u2019s incredibly cost effective and it allows you to address the long-tail revenues that we have lost in the past because of our ground stock model."}', '{"speaker": "James Farley", "utterance": "And the last one is, it\\u2019s lower cost. There is less parts hanging around. We can manage our industrial system and our manufacturing in a leaner way. The question really is, how we maintain it, as you said, as the market improves? And the way we\\u2019re looking at that is not just having a day supply target in the past that we\\u2019ve managed, but actually putting in the infrastructure to maintain or prefer a order-based system. That means we train our system to put in orders. We reward people for putting in orders. We dynamically price for customers so that they are incentivized to keep ordering versus buying off the lot."}', '{"speaker": "James Farley", "utterance": "So it\\u2019s going to be a journey. It\\u2019s been a very rewarding one so far and we\\u2019re just beginning. This is the model we have to go to as most of our business or majority of our business goes battery-electric and digital. It\\u2019s the right loyalty model."}', '{"speaker": "Ryan Brinkman", "utterance": "Okay, great. Thanks. And then just as a quick follow-up to that. It seems that was discussed earlier that product development times are speeding up, maybe particularly with regard to EV. So lighting, for example, seem to come together very fast. Another trend seems to be that automakers are revealing their EVs for a longer period of time before the actual start of production maybe because they\\u2019re so eager to show, then consumers are so clamoring for them. Does that mean that you think that order banks and order in advance might be even more popular with electric vehicles? What are you seeing with regard to that?"}', '{"speaker": "James Farley", "utterance": "The move to a digital product means we have to go to 100% loyalty model. So the reason why you\\u2019re seeing us launch battery-electrics early is very simple, it\\u2019s our Super Bowl ad. Our new Super Bowl ad, our new Detroit Motor Show is our reveal, because it starts the clock on reservations. And you have to do it early enough so your industrial system gets informed by the results of your reservation. That\\u2019s the closed-loop that has to happen. We need to open it early enough so that our industrial system can react to the orders and we don\\u2019t waste money and take advantage of long-tail revenue. And it\\u2019s a more controlled environment than a broadcast media advertising on the Super Bowl."}', '{"speaker": "Ryan Brinkman", "utterance": "Very helpful. Thank you."}', '{"speaker": "Erica", "utterance": "Your next question comes from the line of Rod Lache with Wolfe Research."}', '{"speaker": "Rod Lache", "utterance": "Hi, everybody. I have just two questions. First, you\\u2019ve got a lot of growth that you\\u2019re targeting in BEVs and digital businesses, so it\\u2019s not surprising that we would see some structural cost inflation. What we\\u2019re seeing right now is actually really benign. It\\u2019s $200 million in the quarter, considering what you\\u2019ve got going on. But maybe can you talk a little bit about how we should think about the feathering in of those additional structural costs which presumably come in ahead of the revenue? How should we think about that as we look out to the next year or two?"}', '{"speaker": "John Lawler", "utterance": "Yeah, Rod. We\\u2019ll start to see those come in as we get into \\u201922 and then they will feather into \\u201923 as we continue to ramp our investments and our plan, our priorities, not only in the products to BEVs, but also as we\\u2019re building out our customer-facing technologies, our connectivity, etc. So yeah, you will see that start to come in on a year-over-year basis next year and it will continue in through \\u201923."}', '{"speaker": "Rod Lache", "utterance": "Can you just give us any sort of brackets around what\\u2026 I mean, you did mention that 25% of that EV spending will be expense. But any sort of thoughts on the magnitude of that, what that headwind is?"}', '{"speaker": "John Lawler", "utterance": "I\\u2019m not ready to do that today for \\u201922 and going through \\u201923. We\\u2019re completely targeted on getting to that 8% in 2023, so we\\u2019ll manage it within that. But today, I\\u2019m not ready to talk in that level of detail about it."}', '{"speaker": "Rod Lache", "utterance": "Okay. I was a little surprised about the comment about just 10% volume growth for next year. It seems to me like the Renesas fire and Texas storms alone might have knocked 200,000 units off of your production in Q2. And it wasn\\u2019t too long ago you guys were routinely doing over 700,000 units a quarter. So do you have any thoughts that you might be able to share about when would you be able to get back up to that kind of a level of production? And if so, when should we expect that to happen?"}', '{"speaker": "John Lawler", "utterance": "Right. I think what you\\u2019ll find is that as you look through 2022, the first half, we\\u2019ll have less supply than through the second half. And as I said earlier, we see this mitigating over time. It may extend into 2023, but I would say that we should be back up and running based on what we\\u2019re seeing today, a run rate, the end of next year into \\u201923. And then in 2023, we\\u2019d start to rebuild our inventories. But it\\u2019s dynamic, Rod, and it\\u2019s hard to make a pinpoint call at this point in time. But we wanted to share with you what we\\u2019re seeing, is that we\\u2019re seeing about 10% for next year. And we see that the chip constraint is going to still hit us. It\\u2019s going to still be a factor next year. So we have to keep managing as we are this year."}', '{"speaker": "Rod Lache", "utterance": "Got it. Thank you"}', '{"speaker": "Erica", "utterance": "Your next question comes from the line of Brian Johnson with Barclays."}', '{"speaker": "Brian Johnson", "utterance": "Two questions. First a quick, not quite housekeeping, but definitely balance sheets question. As you restated the dividend, I think at the level, just could you maybe talk us through the investment-grade rating implications and timeline to get there that you can say that you and the board considered when setting that?"}', '{"speaker": "John Lawler", "utterance": "So we\\u2019re going to continue to work and focus on improving our business. Our target is to have an investment-grade balance sheet, but that\\u2019s going to come by improving the business, and you\\u2019re seeing the strength of that come through. And so that\\u2019s what we\\u2019re focused on. What the rating agencies decide to do with our rating, they will manage that, that\\u2019s up to them. What we\\u2019re laser focused on is improving the run rate of the business, improving our performance, improving our overall metrics. And eventually the rating agencies and the ratings will take care of themselves."}', '{"speaker": "Brian Johnson", "utterance": "Okay. Second question. As you think about that 10% volume increase rough guidance, a couple things. One, where do you see fleet sales coming back as you kind of bring that up? And second, are you going to take a different attitude towards fleet sales than in Ford of the past? I remember Don Leclair saying he had two factories making Touristers and rental car companies were about the only buyers of them. But there\\u2019s also really rental cars, but maybe some of the government business that\\u2019s not pleased, it\\u2019s not quite the same. And related to that, as you think about prioritizing production, are there models where you\\u2019re more comfortable you\\u2019ll get good price retention in other models, and I\\u2019m going to pick on like Escape maybe that have a lot of competition in this segment, and as capacity comes back, less likely to hold price say compared to Bronco?"}', '{"speaker": "John Lawler", "utterance": "It\\u2019s interesting because I remember those days when Don probably made that comment about rental fleets and that they were low margin, etc. I think what you\\u2019re finding is business models are changing and the fleet business is evolving just like everything else in our industry. And we see that there could potentially be a positive fleet business where they could make good money. And so we\\u2019re not going to shy away from that if we see that it\\u2019s right for our brand and we think it\\u2019s right for the bottom line. And so we\\u2019re going to continue to look at fleets differently, we\\u2019re going to continue to think about vehicles as a service. And what that potential holds for us if that business model changes, and we\\u2019ll see where that takes us. But we\\u2019re not going to go back to the times where we\\u2019re putting in capacity, we\\u2019re pumping out units, selling them at little to no margin for rental cars. That\\u2019s not going to happen again."}', '{"speaker": "Brian Johnson", "utterance": "how you think about price retention?"}', '{"speaker": "John Lawler", "utterance": "I\\u2019m sorry."}', '{"speaker": "James Farley", "utterance": "It\\u2019s okay. I was just going to say, our fleet business now, now that we\\u2019ve rationalized the company, our fleet business is very strategic for us. It\\u2019s also very profitable. It certainly varies in Europe and North America and China. Different fleet segments have different profitability. The one thing I would ask you to think about is that most of the fleet that matters at Ford is commercial vehicles. And the most important commercial vehicles for us is small, medium-sized businesses, and those are very profitable business for us. For transit, for super duty, that\\u2019s where Ford excels in the fleet business and it\\u2019s smaller fleets. It\\u2019s not big fleet sales."}', '{"speaker": "James Farley", "utterance": "So the texture of this is that we\\u2019re revenue managing the company very carefully. We know the margins by geography even within the country, and we know by distribution channel. So this is a very thoughtful approach for us. But strategically, especially because of the Pro business and its profitability, we want to make sure we\\u2019re a reliable partner with fleet customers. They do business with companies that are reliable. They don\\u2019t come in and out of the market. They do business with companies that have a full range of products, a full range of services. That\\u2019s why Marion is investing in Ford Pro and why we\\u2019re vertically integrating our services. So I think we have a really good profitable fleet business around the world. We look at the margins very carefully, but it\\u2019s strategically very important for the company to be a reliable partner."}', '{"speaker": "Brian Johnson", "utterance": "Okay. And in terms of price retention and how that\\u2019s going to vary across your product line?"}', '{"speaker": "James Farley", "utterance": "Well, John, I think you should answer that one in terms of how we revenue manage in a constrained environment."}', '{"speaker": "John Lawler", "utterance": "Yeah. As you would expect, we\\u2019re very conscientious about the dynamic of the supply and demand and the impact that that has on the pricing. And we look at this on a daily basis, managing our incentives, looking at if we should be taking top-line pricing given the inflationary pressures we\\u2019re seeing. And as we talked about, we\\u2019re not going to go back to the old habits of loading up the dealers with stock and then looking through the push through for sales. We\\u2019re going to focus more on orders coming through online, specific orders to customers being satisfied, understanding what their demands are, simplifying the system. And with all of that, we expect to retain quite a bit of the price. Now, will it mitigate as we go through next year as supply and demand becomes more imbalance and into \\u201923? Yes, but our job is going to be to manage that and retain as much pricing as we can, while providing customers good value for those products. So it\\u2019s something that we look at very closely on a daily basis."}', '{"speaker": "Brian Johnson", "utterance": "Thank you."}', '{"speaker": "James Farley", "utterance": "In the Escape business, we now have another player called Bronco Sport in the segment. It\\u2019s incredibly profitable and people really appreciate the product. We are not in the business of commodity products in that segment anymore. We changed. We made an investment several years ago."}', '{"speaker": "John Lawler", "utterance": "Right."}', '{"speaker": "Erica", "utterance": "Your next question comes from the line of Colin Langan with Wells Fargo."}', '{"speaker": "Colin Langan", "utterance": "Great. Thanks for taking my question. I just want to clarify, I thought your original guidance was that the second half was supposed to be up in volume, 30%. I mean, I\\u2019m not sure if I\\u2019m misreading it, it sounds like Q3 maybe up a bit from Q4. So is that 30% still not accurate? Obviously, it\\u2019s kind of important when you think about the 10% into 2022 what base we\\u2019re going off?"}', '{"speaker": "James Farley", "utterance": "Colin, that\\u2019s a great question. Thank you. Now we did say last quarter that we expected the second half to be up about 30%. Looks like it\\u2019s going to be up somewhere around 15%. And so what you\\u2019re seeing flowed through is the strength we had in the quarter relative to the top-line and other actions that we took relative to cost, etc. So when you look at that walk, that bridge between Q3 to Q4, we expect market factors to be positive. We said we think volume is going to be up sequentially about 10%. We also see a little bit stronger mix continuing. And then of course you\\u2019ll have some production-related costs associated with that. But net-net, market factors net of those costs to produce the increased volumes is going to be positive."}', '{"speaker": "James Farley", "utterance": "What we\\u2019re seeing from a headwind standpoint, if you look at Q3 to Q4, are commodities. We expect that on a quarter-over-quarter basis, they\\u2019re going to be up about $700 million. And if you look at that, so far year-to-date, we\\u2019ve seen about $1.6 billion of commodities hit us. And when you get to the fourth quarter, you can get the cumulative effect that on a year-over-year basis commodities are going to be up about another 1.5 billion in the fourth quarter. So year-over-year, up 1.5 billion, sequentially up $700 million."}', '{"speaker": "James Farley", "utterance": "And then we are going to see some higher warranty costs on a sequential basis in the fourth quarter for things that we have to take care of around extended warranties and a little bit higher coverages. But again, on a year-over-year basis, our warranty will improve in the fourth quarter. In full year, on a year-over-year basis, our warranty we expect to be good about 1.4 billion."}', '{"speaker": "Colin Langan", "utterance": "Got it. That\\u2019s very helpful."}', '{"speaker": "James Farley", "utterance": "So, does that help you with the bridge?"}', '{"speaker": "Colin Langan", "utterance": "Yeah. No, that\\u2019s great. Secondly, in terms of the redesign plan that\\u2019s been out for a while, is India the last major step? I mean, is this going to sort of\\u2026 Is this it or next quarter may be the last time we see these slides? Just curious or is there more still coming?"}', '{"speaker": "James Farley", "utterance": "Well, I think we\\u2019re in good shape for now. Obviously, the acceleration of the BEV business and our ICE assets will be I think the next big transition for the whole industry, not just Ford. But Ford specifically, India\\u2019s really the principal region, country where we have struggled over time. And it\\u2019s really great to see the progress the team is making in India and the very vibrant position we\\u2019ll now have with the new lineup. I\\u2019d just like to highlight the progress in North America for this quarter. John, when is last time we were profitable in South America?"}', '{"speaker": "John Lawler", "utterance": "I believe it was 2013."}', '{"speaker": "James Farley", "utterance": "So let\\u2019s hang that in the air for a second, 2013."}', '{"speaker": "Erica", "utterance": "Your next question comes from the line of Joseph Spak with RBC."}', '{"speaker": "Joseph Spak", "utterance": "Thank you. Maybe just one quick one on the free cash guidance, which I think you maintained despite the EBIT guidance went higher. So is there something going on with working capital? Because you\\u2019re saying you\\u2019re releasing more vehicles, so I wouldn\\u2019t think that would actually be positive factor in the fourth quarter as well as I\\u2019m curious what the offset is?"}', '{"speaker": "John Lawler", "utterance": "Yeah. So what we\\u2019re seeing there is that we\\u2019ve got the EBIT coming in, that improvement, but the less favorable improvement in working capital and timing differences hitting us in the quarter, because we have lower than anticipated volume in the back half of the quarter due to the chip constraints, and so we get with that working capital at the end of the year. So that what\\u2019s happening to us on the free cash flow. So it\\u2019s a timing issue."}', '{"speaker": "Joseph Spak", "utterance": "Okay. And then I want to go back to some of the other announcements you\\u2019ve made over the past couple months and I know you talked again about the spend today. You\\u2019re not spending more than $30 billion, and I appreciate the breakdown you gave of sort of how you\\u2019re spending that. Maybe this is just me, but I actually find it still fairly difficult to track what exactly you\\u2019re spending over the coming years, because I believe some of that has already been spent. So is it possible to maybe just say like of that $30 billion, what\\u2019s being spent, like starting next year through the middle of the decade?"}', '{"speaker": "John Lawler", "utterance": "Of the $30 billion, when you look at the cap, very little\\u2026 We said about half of that was cap. Very little of that has been spent through 2021 relative to the $15 billion, about half of it. Of course, you\\u2019re going to see the expense front-loaded because that\\u2019s primarily the engineering that we have in developing the battery-electric vehicles. And then the direct investment, which is about 25% of it, that\\u2019s for things like the integration of JVs and those types of things. And you saw those announcements this quarter with our plan in BlueOval SK, the battery plant. So that\\u2019s how we\\u2019re going to unfold that BEV spending over time."}', '{"speaker": "Joseph Spak", "utterance": "Okay. That\\u2019s helpful. I appreciate it."}', '{"speaker": "Erica", "utterance": "Your next question comes from the line of Itay Michaeli with Citi."}', '{"speaker": "Itay Michaeli", "utterance": "Great, thanks. Good evening, everybody. Just two quick ones from me. First, is there any update on the BlueCruise deployment, including through OTA, maybe some initial customer feedback?"}', '{"speaker": "James Farley", "utterance": "Great. Thank you. We\\u2019re shipping with Mach-E F-Series now BlueCruise as they leave the factory. And we\\u2019re going to be OTA-ing BlueCruise in the first quarter. We wanted to improve the customer experience, so we pushed it back in terms of an OTA because we wanted it to be much simpler for the customer than was originally planned, and that takes a little planning to consolidate. Often these Level 2 systems require multiple updates to the car. We wanted it to be very simple. That took a little bit more work on our team\\u2019s part. And so it\\u2019s available as we ship products now. And as an OTA, it will be in the first quarter and it will be a lot simpler to use and get that OTA and update for the customer than it was originally planned. Does that answer your question?"}', '{"speaker": "Itay Michaeli", "utterance": "Yeah. Thank you, Jim. That\\u2019s very, very helpful. And then maybe just a super quick follow-up, just a point of clarification. Thank you for the 2022 initial indications. In the release, you mentioned you expect to build on the strong performance in 2021. I am just curious if we should interpret that as you expect to grow EBIT-adjusted year-over-year in 2022?"}', '{"speaker": "John Lawler", "utterance": "Yeah. We\\u2019re not going to give a number at this point, Itay. But what we\\u2019re saying is that the strength of our new product lineup, our high volume nameplates and the strength of what we\\u2019re seeing from Mach-E, as Jim said, we think there\\u2019s about demand for 200,000 units. We\\u2019ve got the Bronco, Maverick, E-Transit, and F-150 Lightning are coming. It\\u2019s the best lineup that we\\u2019ve had. And so that\\u2019s going to be a tailwind for us for sure as we go into next year."}', '{"speaker": "John Lawler", "utterance": "And you\\u2019re seeing that come through this year, and we\\u2019re going to build on that. But we\\u2019re also going to have to manage the headwinds that we\\u2019ve talked about and the other puts and takes. But what we can tell you is we are laser-focused on getting to the 8% target in 2023. And so we will manage into next year. These are the types of things we\\u2019re seeing from a puts-and-takes standpoint, strength, the tailwinds and then the headwinds. And we will manage that next year and we\\u2019ll be on the path next year towards our 8% target in 2022."}', '{"speaker": "Itay Michaeli", "utterance": "Great. That\\u2019s super helpful. Thank you."}', '{"speaker": "Erica", "utterance": "Your next question comes from the line of Emmanuel Rosner with Deutsche Bank."}', '{"speaker": "Emmanuel Rosner", "utterance": "Hi. Thank you very much, and good evening. Two questions please-"}', '{"speaker": "James Farley", "utterance": "Hi. Good evening."}', '{"speaker": "Emmanuel Rosner", "utterance": "Hi. Two questions. The first one, very pleased to see beyond target for this 8% margins by 2023, big focus of the company and clearly showing some progress there. How should we think about the impact on this from the margins on your electric vehicles? Obviously, the scale takes some time to build up, but you are going through that right now, and probably has some level of visibility. So what\\u2019s are you really the thinking around trajectory for margins on some of your EVs? And to what extent it will or will not impact overall company margins? And potentially, how do we think about it beyond 2023?"}', '{"speaker": "James Farley", "utterance": "Thanks. Such an important question for the company. I\\u2019d like John to comment on the margins. Right now we have three high volume, very well-accepted battery-electric vehicles on our hands; Mach-E, E-Transit and the F-150 Lightning. So the way we look at it is, we want to grow this business really fast. Just the Mach-E demand itself we think is 200,000 units, that do not include the Lightning or the E-Transit."}', '{"speaker": "James Farley", "utterance": "So our first job is of course post job one customer experience improvement, post job one simplification and improvement of the cost of the vehicle, and post job one quality improvements using the data off the vehicles. But perhaps our biggest job in my opinion is to break the constraints we have in manufacturing and our supply chain so we can get these products out to these customers."}', '{"speaker": "James Farley", "utterance": "That post job one orientation is quite different than how we historically looked at the ICE business where we wait to a minor change or something later to make those changes. The constraint for Mach-E right now is batteries. We think we can break some of those constraints by working creatively with our China team and get our batteries from China. So stay tuned. I\\u2019ll ask John to comment on the margins."}', '{"speaker": "John Lawler", "utterance": "All right. Well, as we talked about last quarter, Mach-E is EBIT profitable today, but we also know that the margins are not as strong as our ICE margins. And so we\\u2019re working on that. Over time, we expect as we scale, as you said, and as the technology costs come down, we will grow those margins. And ultimately, we do expect with these connected vehicles, these connected bands that the profit margins will be better than what we\\u2019re seeing on ICE today, but that\\u2019s over time."}', '{"speaker": "Emmanuel Rosner", "utterance": "Thank you. And then my second question was on Argo. So very encouraged to see that you would like to encourage them as supportive of accessing the public markets. How do you envisage the future relationship between Ford and Argo to be? How important is that going to be as part of your overall business model?"}', '{"speaker": "James Farley", "utterance": "Mission critical. For us to truly disrupt personal ownership, we have to democratize shared mobility. And the self-driving and mobility in the driven world are absolutely mission critical for the company and disrupt itself. I am really proud of the team\\u2019s progress. It\\u2019s different than our competitors in the space. We\\u2019ve gone to the most difficult miles in four or five different cities."}', '{"speaker": "James Farley", "utterance": "Our mapping, our SDS deployment and the algorithms are built to be scaled for production deployment. So we\\u2019re not going to a small market area and easy miles like others, we\\u2019re taking on the toughest problems now and building our capability for scaling quickly. I think that\\u2019s always been our approach. The relationship with Argo and us and Volkswagen is very close, but we do see us moving into more of a production mode now and we\\u2019re really ready for that. We think this will take more capital and little more time, and we think the access to public capital is really mission critical for our journey."}', '{"speaker": "Emmanuel Rosner", "utterance": "Thank you."}', '{"speaker": "Erica", "utterance": "Your final question comes from the line of Adam Jonas with Morgan Stanley."}', '{"speaker": "Speaker 15", "utterance": "Ford."}', '{"speaker": "Adam Jonas", "utterance": "There you go. He\\u2019s a Ford fan. We got a Ford fan here. Nobody has I think explained, as lucidly and clearly the in Ford Pro stuff the way you have. So you get kudos for that. You\\u2019re seeing some of your competitors make some critical investments in the downstream to make up for that customer experience; Volkswagen buying your car, and some start-ups, including ones that you know pretty well kind of owning that physical part of it. Is there any gap in your strategy as you go always on and really the Ford service\\u2026 Is there anything that you want to vertically integrate to either the Ford umbrella, or is working with the franchisers and the third parties sufficient given this big change in? Thanks."}', '{"speaker": "James Farley", "utterance": "Great question. Our philosophy is different. We think partnership on the demand layers for autonomy and pre-autonomy is mission critical for our always-on strategy. Are there pieces missing that we\\u2019re working really hard on? You bet you. We\\u2019re not going to talk about them today, though."}', '{"speaker": "Adam Jonas", "utterance": "All right. Follow-up for you and your dealers. They\\u2019re crushing it. Some people say they\\u2019re gouging, that might be unfair because they\\u2019re paying high prices, too, for the vehicles, but we\\u2019re starting to see four handles, five handles, six-handled GPUs all in, way big chunk at the price of a car, Jim."}', '{"speaker": "Adam Jonas", "utterance": "think you\\u2019re making too much because I know that\\u2019s a grouchy question. Do you think as far as that, that you can for Ford in the consumer?"}', '{"speaker": "James Farley", "utterance": "Well, this is also really important. First of all, I would say the heart and soul of Ford\\u2019s strategy is our commercial business. In that business where vehicles are highly utilized, our dealer network is one of our most important advantages versus the new competitors. I will give you some statistics. We have 650 dedicated commercial, mostly service centers in the United States and 850 transit centers across Europe. That will take a lot of time and a lot of money for someone to recreate."}', '{"speaker": "James Farley", "utterance": "Every one of those dealers has multiple bodybuilders that can call design for those trades and those vocations for our Ford vehicles. So the dealer body is not only important for the after-sales experience and making sure those vehicles can be serviced, but it\\u2019s also mission critical for the outfit of those products. So the dealer network is absolutely strategically critical for our leadership and maintaining that as we go to digital-connected vehicles for our commercial customers."}', '{"speaker": "James Farley", "utterance": "There\\u2019s no doubt that many customers want a three or four-click, very easy service experience on the retail side, and we\\u2019re working really carefully on that, including a simple e-commerce platform. And actually in China, our BEV business already has 25 direct stores by the end of this year. So we\\u2019re starting to experiment. I think our dealers has served us really well. I\\u2019m very proud of them. I\\u2019m especially proud of our commercial dealers and we are very vigilant. You can imagine, I get lots of emails every day about transaction prices from customers on our hottest products. And we all feel obligated to represent the brand professionally for our customers."}', '{"speaker": "Adam Jonas", "utterance": "Thanks, Jim."}', '{"speaker": "Erica", "utterance": "This concludes the Ford Motor Company third-quarter 2021 earnings conference call. Thank you for your participation. You may now disconnect."}']
}
res = requests.post(
"https://solutions.forefront.ai/v1/organization/{TEAM_ID}/summarize",
json=body,
headers=headers
)
data = res.json()
After sending the dialogue to the Summarize API, the original 1 hour and 6 minute transcript is summarized to (at compression_level = 5):
Summary (4,318 characters)

Summarize text

An example use case for summarizing text would be to summarize a research paper. Take the following paper, Language Models are Few-shot Learners, and let's summarize the Introduction.
Introduction to Language Models are Few-Shot Learners (10,612 characters)
We can summarize all of the text in the following API call:
import requests
import json
headers = {
"Authorization": "Bearer {API_KEY}",
"Content-Type": "application/json"
}
body = {
"compression_level": 3,
"text": "Recent years have featured a trend towards pre-trained language representations in NLP systems, applied in increasingly flexible and task-agnostic ways for downstream transfer. First, single-layer representations were learned using word vectors [MCCD13, PSM14] and fed to task-specific architectures, then RNNs with multiple layers of representations and contextual state were used to form stronger representations [DL15, MBXS17, PNZtY18] (though still applied to task-specific architectures), and more recently pre-trained recurrent or transformer language models [VSP+17] have been directly fine-tuned, entirely removing the need for task-specific architectures [RNSS18, DCLT18, HR18]. This last paradigm has led to substantial progress on many challenging NLP tasks such as reading comprehension, question answering, textual entailment, and many others, and has continued to advance based on new architectures and algorithms [RSR+19, LOG+19, YDY+19, LCG+19]. However, a major limitation to this approach is that while the architecture is task-agnostic, there is still a need for task-specific datasets and task-specific fine-tuning: to achieve strong performance on a desired task typically requires fine-tuning on a dataset of thousands to hundreds of thousands of examples specific to that task. Removing this limitation would be desirable, for several reasons. First, from a practical perspective, the need for a large dataset of labeled examples for every new task limits the applicability of language models. There exists a very wide range of possible useful language tasks, encompassing anything from correcting grammar, to generating examples of an abstract concept, to critiquing a short story. For many of these tasks it is difficult to collect a large supervised training dataset, especially when the process must be repeated for every new task. Second, the potential to exploit spurious correlations in training data fundamentally grows with the expressiveness of the model and the narrowness of the training distribution. This can create problems for the pre-training plus fine-tuning paradigm, where models are designed to be large to absorb information during pre-training, but are then fine-tuned on very narrow task distributions. For instance [HLW+20] observe that larger models do not necessarily generalize better out-of-distribution. There is evidence that suggests that the generalization achieved under this paradigm can be poor because the model is overly specific to the training distribution and does not generalize well outside it [YdC+19, MPL19]. Thus, the performance of fine-tuned models on specific benchmarks, even when it is nominally at human-level, may exaggerate actual performance on the underlying task [GSL+18, NK19]. Third, humans do not require large supervised datasets to learn most language tasks – a brief directive in natural language (e.g. “please tell me if this sentence describes something happy or something sad”) or at most a tiny number of demonstrations (e.g. “here are two examples of people acting brave; please give a third example of bravery”) is often sufficient to enable a human to perform a new task to at least a reasonable degree of competence. Aside from pointing to a conceptual limitation in our current NLP techniques, this adaptability has practical advantages – it allows humans to seamlessly mix together or switch between many tasks and skills, for example performing addition during a lengthy dialogue. To be broadly useful, we would someday like our NLP systems to have this same fluidity and generality. One potential route towards addressing these issues is meta-learning1 – which in the context of language models means the model develops a broad set of skills and pattern recognition abilities at training time, and then uses those abilities at inference time to rapidly adapt to or recognize the desired task (illustrated in Figure 1.1). Recent work [RWC+19] attempts to do this via what we call “in-context learning”, using the text input of a pretrained language model as a form of task specification: the model is conditioned on a natural language instruction and/or a few demonstrations of the task and is then expected to complete further instances of the task simply by predicting what comes next. While it has shown some initial promise, this approach still achieves results far inferior to fine-tuning – for example [RWC+19] achieves only 4% on Natural Questions, and even its 55 F1 CoQa result is now more than 35 points behind the state of the art. Meta-learning clearly requires substantial improvement in order to be viable as a practical method of solving language tasks. Another recent trend in language modeling may offer a way forward. In recent years the capacity of transformer language models has increased substantially, from 100 million parameters [RNSS18], to 300 million parameters [DCLT18], to 1.5 billion parameters [RWC+19], to 8 billion parameters [SPP+19], 11 billion parameters [RSR+19], and finally 17 billion parameters [Tur20]. Each increase has brought improvements in text synthesis and/or downstream NLP tasks, and there is evidence suggesting that log loss, which correlates well with many downstream tasks, follows a smooth trend of improvement with scale [KMH+20]. Since in-context learning involves absorbing many skills and tasks within the parameters of the model, it is plausible that in-context learning abilities might show similarly strong gains with scale. In this paper, we test this hypothesis by training a 175 billion parameter autoregressive language model, which we call GPT-3, and measuring its in-context learning abilities. Specifically, we evaluate GPT-3 on over two dozen NLP datasets, as well as several novel tasks designed to test rapid adaptation to tasks unlikely to be directly contained in the training set. For each task, we evaluate GPT-3 under 3 conditions: (a) “few-shot learning”, or in-context learning where we allow as many demonstrations as will fit into the model’s context window (typically 10 to 100), (b) “one-shot learning”, where we allow only one demonstration, and (c) “zero-shot” learning, where no demonstrations are allowed and only an instruction in natural language is given to the model. GPT-3 could also in principle be evaluated in the traditional fine-tuning setting, but we leave this to future work. Figure 1.2 illustrates the conditions we study, and shows few-shot learning of a simple task requiring the model to remove extraneous symbols from a word. Model performance improves with the addition of a natural language task description, and with the number of examples in the model’s context, K. Few-shot learning also improves dramatically with model size. Though the results in this case are particularly striking, the general trends with both model size and number of examples in-context hold for most tasks we study. We emphasize that these “learning” curves involve no gradient updates or fine-tuning, just increasing numbers of demonstrations given as conditioning. Broadly, on NLP tasks GPT-3 achieves promising results in the zero-shot and one-shot settings, and in the the few-shot setting is sometimes competitive with or even occasionally surpasses state-of-the-art (despite state-of-the-art being held by fine-tuned models). For example, GPT-3 achieves 81.5 F1 on CoQA in the zero-shot setting, 84.0 F1 on CoQA in the one-shot setting, 85.0 F1 in the few-shot setting. Similarly, GPT-3 achieves 64.3% accuracy on TriviaQA in the zero-shot setting, 68.0% in the one-shot setting, and 71.2% in the few-shot setting, the last of which is state-of-the-art relative to fine-tuned models operating in the same closed-book setting. GPT-3 also displays one-shot and few-shot proficiency at tasks designed to test rapid adaption or on-the-fly reasoning, which include unscrambling words, performing arithmetic, and using novel words in a sentence after seeing them defined only once. We also show that in the few-shot setting, GPT-3 can generate synthetic news articles which human evaluators have difficulty distinguishing from human-generated articles. At the same time, we also find some tasks on which few-shot performance struggles, even at the scale of GPT-3. This includes natural language inference tasks like the ANLI dataset, and some reading comprehension datasets like RACE or QuAC. By presenting a broad characterization of GPT-3’s strengths and weaknesses, including these limitations, we hope to stimulate study of few-shot learning in language models and draw attention to where progress is most needed. A heuristic sense of the overall results can be seen in Figure 1.3, which aggregates the various tasks (though it should not be seen as a rigorous or meaningful benchmark in itself). We also undertake a systematic study of “data contamination” – a growing problem when training high capacity models on datasets such as Common Crawl, which can potentially include content from test datasets simply because such content often exists on the web. In this paper we develop systematic tools to measure data contamination and quantify its distorting effects. Although we find that data contamination has a minimal effect on GPT-3’s performance on most datasets, we do identify a few datasets where it could be inflating results, and we either do not report results on these datasets or we note them with an asterisk, depending on the severity. In addition to all the above, we also train a series of smaller models (ranging from 125 million parameters to 13 billion parameters) in order to compare their performance to GPT-3 in the zero, one and few-shot settings. Broadly, for most tasks we find relatively smooth scaling with model capacity in all three settings; one notable pattern is that the gap between zero-, one-, and few-shot performance often grows with model capacity, perhaps suggesting that larger models are more proficient meta-learners. Finally, given the broad spectrum of capabilities displayed by GPT-3, we discuss concerns about bias, fairness, and broader societal impacts, and attempt a preliminary analysis of GPT-3’s characteristics in this regard. The remainder of this paper is organized as follows. In Section 2, we describe our approach and methods for training GPT-3 and evaluating it. Section 3 presents results on the full range of tasks in the zero-, one- and few-shot settings. Section 4 addresses questions of data contamination (train-test overlap). Section 5 discusses limitations of GPT-3. Section 6 discusses broader impacts. Section 7 reviews related work and Section 8 concludes."
}
res = requests.post(
"https://solutions.forefront.ai/v1/organization/{TEAM_ID}/summarize",
json=body,
headers=headers
)
data = res.json()
After sending the text to the Summarize API, the original 10,612 characters is summarized to 2,857 characters (at compression_level = 3):
Summary (2,857 characters)
Last modified 19d ago
Copy link
On this page
Create summary
post
Create summary
Example request
Example use cases