Aug 26, 2025 · 47m · a16z
How Scale AI is Pioneering the Future of Work
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the a16z Enterprise podcast, Ben Scharfstein, Head of Product for Enterprise Solutions at Scale AI, joins host Joe Schmidt to discuss how forward-deployed engineering teams, custom enterprise AI implementations, and internal AI automation are transforming software delivery and creating defensible market moats.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The host as informed peer 4.2 Guest teaching 4.1 Guest disagreement 0.9 The host pushing back 0.5
speaking balance: gold is the host, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 7:35 Rejection of Low-Margin Services Stigma
Ben rejects the premise that services-led motions are inherently poor businesses, citing Palantir's stellar performance and explaining how AI agents transform internal knowledge work.
Hardest push from the host ▶ 27:34 Testing Down-Market Economic FloorJoe presses Ben directly on whether forward-deployed motions are viable for smaller SMBs or if an absolute contract size floor exists.
Biggest teaching moment ▶ 12:35 Deconstructing Moats via Helmer's 7 PowersBen provides a masterclass on strategic defensibility, explaining why software alone offers no moat and showing how converting human SOPs into AI-ready data assets generates network effects.
The host holds their own ▶ 38:31 Trading Margin for Moat ThesisJoe demonstrates deep strategic expertise when queried by Ben, delivering an extended analysis on why startups should sacrifice short-term gross margins to lock down mission-critical enterprise workflow layers.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The host as informed peer | Guest teaching | Guest disagreement | The host pushing back | Why |
|---|---|---|---|---|---|---|
| Scale AI's Enterprise Application Business Overview | 2 | 3 | 0 | 0 | Joe welcomes Ben and prompts him to explain Scale AI's application business. Ben clarifies Scale's dual focus between model data labeling and building custom enterprise AI agents. | |
| Enterprise AI Adoption: From Pilots to Production | 3 | 4 | 1 | 0 | Joe highlights Scale's rapid nine-figure revenue growth in applications and asks what triggered the shift. Ben explains that enterprise adoption lagged Twitter hype by 18 months due to change management and data integration hurdles. | |
| Vertical AI vs. Custom Enterprise Solutions | 4 | 4 | 1 | 1 | Joe compares Scale's approach to vertical AI plays like Harvey and Decagon. Ben explains that top-tier enterprises reject generic peer averages in favor of customized solutions that protect their secret sauce. | |
| Scaling Services with AI Agents and Internal Automation | 4 | 5 | 2 | 1 | Joe questions whether services-led models scale linearly with headcount. Ben counters by pointing to Palantir's success and explaining how internal AI agents augment knowledge work to defy traditional services economics. | |
| Tactical Guidance for Startups: Forward-Deployed Wedges | 3 | 5 | 1 | 0 | Joe asks for tactical guidance for AI startups seeking enterprise entry points. Ben outlines how forward-deployed engineering serves as a wedge into fortune companies where standard SaaS configuration falls short. | |
| Building Moats in AI: Network Effects and Data Assets | 3 | 5 | 1 | 0 | Joe probes into how startups can build durable moats in a rapidly shifting AI ecosystem. Ben cites Helmer's 7 Powers to argue that software is not a moat, whereas turning human SOPs into proprietary AI data assets creates lasting defensibility. | |
| Unpacking Forward-Deployed Roles: FDE, FDMLE, and FDPM | 4 | 5 | 1 | 0 | Joe references his published work on forward-deployed roles being Silicon Valley's hottest trend. Ben breaks down the distinct operational mandates of FDEs, FDMLEs, and FDPMs. | |
| Overcoming "Schlep Blindness" and Managing Scope Creep | 4 | 5 | 1 | 1 | Joe asks how teams navigate scope creep when enterprise clients request legacy integrations. Ben invokes Paul Graham's 'schlep blindness' concept, explaining that tackling unsexy integration work is necessary to solve C-suite stock-moving problems. | |
| Customer Strategy: Enterprise Bets vs. Agile Design Partners | 3 | 4 | 0 | 0 | Joe asks how to find the 'Goldilocks zone' in enterprise customer sizing. Ben details Scale's portfolio strategy of pairing long-cycle Fortune 500 accounts with fast-moving design partners. | |
| Applying Forward-Deployed Motions Down-Market and Economic Floors | 4 | 4 | 1 | 1 | Joe asks if forward-deployed motions can work down-market for SMBs and pushes on ACV thresholds. Ben clarifies that deploying forward-deployed engineering teams against small $20k contracts is economically unviable. | |
| Long-Term Horizon of Forward-Deployed Teams and AI Integrators | 4 | 5 | 1 | 0 | Joe questions if forward-deployed teams will eventually transition to external systems integrators like Salesforce's ecosystem did. Ben predicts a 5-10 year window before coding agents automate integration work and details key hiring traits for FDPMs. | |
| Sales Collaboration and Incentive Alignment for Forward-Deployed Teams | 8 | 2 | 1 | 2 | Ben turns the tables and asks Joe about his post 'Trading Margin for Moat'. Joe takes center stage, explaining his venture thesis on why sacrificing short-term gross margins to capture core workflow context creates defensible long-term value. | |
| Monetizing Implementation and Maintaining Strategic Focus | 7 | 3 | 0 | 1 | Joe highlights a critical mistake startups make by giving away implementation for free instead of using it for ACV discovery. Ben agrees and adds that companies must avoid vanity revenue metrics. | |
| Hot Takes: Foundation Model Blockbusters and Revenue Discipline | 6 | 3 | 1 | 0 | The pair exchange hot takes to conclude the episode. Ben compares foundation model labs to movie studios producing short-payback blockbusters, while Joe argues revenue teams must say 'no' more often to maintain strategic focus. |
Statements from this episode (31)
Scale AI's enterprise applications make up half of its total business
“The other half of our business is the application business”
Scharfstein: Scale AI's enterprise business found PMF in past 12-18 months
“There's been an enterprise team at scale for a long time, but really just in the past 12 to 18 months has really taken off as something that's found kind of product market fit.”
Scharfstein: Rapid AI shifts make custom enterprise builds superior to static products
“That's because the industry is changing a lot. Every day, ah, it changes, you know, GPT-V comes out, and it's different than, you know, it was three months ago, and so what we need to actually build in product changes, and so we've just said, you know what, th…”
Ben Scharfstein says enterprise AI lags public state-of-the-art by 18 months
“And I would say it was probably like an 18 month lag between what you see on Twitter and what's the state of the art and what is actually working in enterprises.”
Scharfstein: Enterprise AI delay is caused by integration and UX, not models
“And I think the interesting thing is it's not because of model capabilities. It's because of change management and it's because of integrations, oftentimes with data and also just kind of understanding like what is the UX? What's the paradigm? The security, al…”
Ben Scharfstein says top enterprises reject off-the-shelf vertical AI products
“And for those companies, they actually don't want the average of their peers. They don't want, you know, the thing that everyone else is getting, what's in distribution for the models. They don't want what's been productized for everyone else because they have…”
Ben Scharfstein predicts AI agents will drive software writing costs to zero
“One of the things that we really believe is the cost of writing software is going to zero in that world.”
Scharfstein: AI aims to augment human work, not replace software
“What we're moving in AI is that we're not trying to replace software. We're trying to augment or automate human work.”
Scharfstein: Enterprise AI shift demands forward-deployed engineering for custom software
“And what we're seeing now with this platform shift is enterprises still need all of that customization. They need all of the feature set, but it doesn't exist in the software and doesn't exist anywhere. And so that's the value of this forward deployed motion i…”
Scharfstein: Forward-deployed engineering must build durable software, not consulting
“And the key is that it should be the customization and a wedge into installing your software that is durable over time. It's not worth just doing this as, you know, there may be great consulting businesses to build, but that's not, you know, the pot of gold.”
Scharfstein: Enterprise AI companies must become systems of intelligence
“I think that's the key thing is you want to become a system of record system of work and eventually become a system of intelligence where you actually do the work.”
Ben Scharfstein argues that software code itself is not a defensible moat
“Interestingly, software was never a moat. It was never something that at least, you know, he thought was leading to differentiated returns. I think that's still true, which is that You know, the software itself is not emote.”
Scale AI employs over 100,000 global contributors for AI data tasks
“We have contributors, you know, a 100,000 plus contributors around the world That do these tasks.”
Scharfstein: Forward-deployed engineering must feed custom client fixes back into platform core
“Really the mandate of the forward deployed engineer is to do that. You do need to say, okay, we have 60% out of the box, but that last 40%, maybe it's a data integration, it's a visualization, it's an agent that we haven't built yet. You do need to do that wor…”
Scharfstein: Scale AI splits forward-deployed teams into software, ML, and product
“At scale, we break down the forward deployed role, not just a forward deployed engineers, which, you know, Palantir made famous, but we also have forward deployed product. That's the team that I lead and forward deployed machine learning engineers or an applie…”
Scharfstein: Solving end-to-end enterprise problems unlocks 10x larger contracts
“Like your job is not to build a product. Your job is to solve a problem. And that is what enterprises are going to expect. That is what's going to unlock 10 X bigger contracts with them.”
Ben Scharfstein estimates enterprise AI solutions are 70 percent unglamorous integration work
“What does the end-to-end solution look like? Which only 30% of it might involve AI, and the rest is that schlep that you just have to do to have an impact.”
Scharfstein: Scale AI allows large enterprise clients to retain their IP
“Oftentimes the big customers want to retain the IP. We're very happy to allow them to retain that IP because, you know, it's core to their business.”
Scharfstein: Enterprise AI startups should not compete with capable internal teams
“You don't want to be competing against these internal teams, especially really capable internal teams.”
Scharfstein: Startups must retain IP and standardize features when serving small clients
“I think the smaller the company and the smaller the customer, the more that you need to retain all of the IP and build it back into the platform.”
Scharfstein: Forward-deployed engineering models are not scalable outside enterprise sales
“If you're not selling to enterprises, it's difficult to think about this as a scalable business model.”
Scharfstein: In-house forward-deployed engineering teams have a 5-to-10-year lifespan
“I think that this idea of having this forward deployed in house, it has a five to 10 year life because we just have so much software left to build.”
Ben Scharfstein predicts AI agents will replace systems integrators within a decade
“It's hard for me to think that in 10 years, there's going to be a super robust systems integrator ecosystem. Probably agents will just do it.”
Scharfstein: Difficult 10x engineers belong on platform teams, not client teams
“That's like very hard to work with, but they're a 10 X engineer. Those people should be on your platform team and not on your forward deploy team.”
Scharfstein: Forward-deployed engineering roles are a factory for founders
“I think it's like a factory for founders, whether or not it's in the product role or it's in the Ford deployed engineering role or AI role, because these are people that are just doing customer discovery while they're building, which is like what being a found…”
Scharfstein: Spending 3 days on-site accelerates enterprise AI delivery by 3 weeks
“Sitting with the customer, you know, flying to six hours away and saying, I'm going to spend three days with you is invaluable. Speeds you up three weeks.”
Scharfstein: Forward-deployed deployment goals must remain separate from sales incentives
“In terms of compensation structure, I'm not an expert, but I will say that I think in the forward deployed motion, the goal is to build a repeatable process through your customers and having that orthogonal to your go-to-market incentives is important because …”
Joe Schmidt says targeting 85 percent AI gross margins is a mistake
“I think that's a tremendous mistake. Like I think it's actually probably the biggest mistake you can make right now.”
Scale AI's Ben Scharfstein: AI startups must avoid vanity revenue metrics when delivering services
“One is I think you just have to be very honest and very sober about your revenue when you do that. And you say, Hey, like we were able to chart, get to ten million dollars in revenue, but it's, Consulting-ish revenue. Our repeatable aspect is two million, and …”
Schmidt: Not charging enterprise clients for AI implementation is a mistake
“I think that's a huge mistake that a lot of the early stage companies can make is you go out there and again, like, you know, you're like, I don't want to charge for implementation. Like these big companies, they won't like that. Like, you know, they just want…”
Ben Scharfstein says AI foundation model labs operate like movie studios
“I think that the, I would say my hottest take is, I think the right way to think about foundation labs is that they're like movie studios. And what I mean by that is they invest a ton of money in blockbusters that have a relatively short time spend to pay them…”