Forward-Deployed Engineers (FDEs) - AI's New ROI Battleground

**Ray Rike** (0:08)
Welcome to this week's AI to ROI The Big Story episode. I'm Ray Rike, founder and CEO of Benchmarkit, co-host of the Metrics Brothers Podcast, and I'm joined today by my co-host, Peter Buchanan.

**Peter Buchanan** (0:26)
Yeah, I'm Peter Buchanan. The managing partner of DuPlan. Today's episode is dedicated to the FDE revolution. That's the forward deployed engineer revolution, and why so much investment is being made in the role across frontier model companies, hyperscalers, and consulting companies. But before we dive in, let's hear a word from our sponsor.
Okay, let's start today's episode with a number that should worry every enterprise buyer.
So AI spending is accelerating faster than almost anything ever tracked. We certainly track it every week in the newsletter, but the ROI story hasn't caught up. So companies are basically running out of budget before they run out of enthusiasm. And this week's story is about a proposed fix, which is turning AI labs into professional services firms. So Ray, what's that all about?

**Ray Rike** (1:28)
Well, I think someone was just calling me. I think it was a CFO, seeing if they could get their AI ROI in place. But seriously, one of the things that's really interesting to me is that everybody is finally looking at their expense trends around AI, and how do I manage it? And the other thing they're really questioning is, what's the return on investment I'm getting for these AI investments that I'm making, especially from the CFO? So I think this is why we're seeing this huge movement by a cross section of vendors to help their enterprise customers successfully deploy and drive benefit from this. I mean, let's look at OpenAI. They announced a $4 billion consulting joint venture back in May. The same day they announced this joint venture, which we're going to dive into a little bit more later. The India's Nifty IT index, a lot of those IT service and consulting companies fell 3.7% and kept declining for four straight days. Infosys, one of the big three India-based consulting firms, hit a 52-week low.
Tata Consulting Services' stock dropped to its lowest point since 2020
So the market treated this announcement by OpenAI getting into the FDE and consulting business as an extinction event for professional service firms. Which by the way, already do a hell of a lot of AI consulting.

**Peter Buchanan** (3:05)
They do, and some of those incumbents actually invested in the new OpenAI venture. So you fast forward two months now, we're in mid-July, we now have five of these ventures from the major AI labs and the top three hyperscalers. So they're based on a consulting approach that employs forward deployed engineers. So I used that word before. That sounds different, but maybe it's not so different. It's just sort of catching up with all these companies. So what's an FDE and where did it come from?

**Ray Rike** (3:39)
Peter, everything old is new again.

**Peter Buchanan** (3:42)
Exactly.

**Ray Rike** (3:42)
So the term forward deployed engineer is new and we're going to talk about its starting point. But I dare to say that I ran a field deployed engineering organization early in my career, back in the early nineties. And we called it a tech center of excellence or a technical development center or field-based technical development center. It's where I had engineers responsible for developing customer specific functionality on top of our core platform. And then we were trying to backward integrate parts of what they deployed to a customer and to the core product. But that's ancient history. Let's get into the 21st century here. In today's world, Palantir, by the way, most people don't understand or know that Palantir is 20 plus years old. They're credited with building this forward deployed engineer function in 2005 because they were trying to solve a problem for some of their early customers. The CIA, the NSA, the US. Army, that traditional consultants who really didn't understand AI and the Palantir technology could address. And the idea was pretty simple. Go ahead and put highly technical engineering teams directly inside of the customer's environment to configure, I would say to enhance and configure Palantir's products. Back then, something called Gotham, Foundry, or AIP, against the customer's actual data, data structures and processes, workflows, and then ship something that works before, you talk about how great my product roadmap is, we actually deployed there in real time. In fact, I think it was a gentleman named Bob McGrew, he was an early Palantir executive and he now works at OpenAI.
He described the FDE's job as product discovery. Go to the customer's site, take the product as it is, and develop and fill the gap between what it does today and what the customer actually needs.