We are natural cynics about bullshit new innovation language. Our industry barfs out fresh chunks in every cycle. Repeated exposure has built some immunity. So, when a phrase spreads as fast as "forward deployed" has this year, the reflex is to roll our eyes and wait for it to pass. First it described an engineer. Then the designer. Recently, I have even heard the term "forward deployed creative" whispered on LinkedIn.
It began at Palantir, as a way to describe engineers who embed inside a customer's business to make the technology work in their messy reality rather than handing it over and walking away. In the past year it has become the explicit strategy of the biggest AI labs and the oldest consulting firms. OpenAI, Anthropic, Google, and Cohere have each built dedicated forward-deployed organisations; Accenture, Deloitte, and EY have each launched their own version. And now it is being bolted onto everything. Forward deployed anything.
Part of me still expects it to pass. This is how good ideas get hollowed out: a precise term spreads faster than the practice behind it until it means nothing. We're still traumatised by what happened to "agile", "lean" and "digital transformation". Put "forward deployed" in front of "designer" and it could just be the next casualty.
But. This one feels different. What we're watching is a kind of co-evolution. Machine intelligence reshapes how people work together; that reshapes what we ask the machines to be; and round it goes, faster each turn. Like two jazz musicians locked in, trading phrases: one plays, the other answers and raises it. The tune neither of them wrote gets quicker and stranger with every pass. We make the tools, and the tools make us.
The canonical forward deployed engineer was framed as a technical role: embed, integrate, own the outcome, ship the thing. But, as ever, the engineering was never the hard part. The hard part is reading the business, working through the messy human stuff, noticing what the technology now makes possible that nobody even knew they needed, and figuring out what is worth building at all. That is design and strategy work. And, increasingly, research too. All of it in the same head, at the same time.
Underneath all this, the disciplines themselves are dissolving into one another, and so is the way the work moves. It used to run as a line with a few loops in it. Now it's a crazy squiggle. My friend Chris Downs, who founded the studio Normally, puts it brilliantly: the work has stopped feeling like pushing water through pipes, with a hard edge at every handoff, and started to feel like swimming in it. If the work no longer breaks into stages, the person doing it stops breaking into separate roles. So the truer name is "forward deployed designer", or more precisely "forward deployed design-engineer". The role no longer sits in a single lane, because the lanes have gone.
This has been taking shape for years, not as a job title but as a pattern in the work. Plenty of people have felt it. There's a thoughtful conversation already running, from David Hoang's embedded squad to Jakob Nielsen's full-dress definition. The one who rings truest to me is Ron Bronson, who ran a 40-person design division at 18F for four years and was doing this before it had a name. His teams worked the way I think the role actually wants to: designers who could code, researchers who could do content and architecture, service designers who could lead and PM, every designer a strategist on some level. Not specialists handing off to specialists. When someone who ran it at that scale lands in the same place, that's the signal worth trusting. Something is forming, and it is bigger than a name.
Most accounts retain the handoff, though. From the first attempts to name the role to Nielsen's careful definition, the shape is the same: the designer works out what to build, and someone else builds it. Nielsen makes it explicit, arguing for the designer and the engineer as two specialists in symbiosis. But that seam is the very thing disappearing. The whole point of the design-engineer is that working out and making have collapsed into one act, in one pair of hands. Not a tidier split. The end of it.
What makes this possible now is partly that the tools have closed the distance between an idea and a working version of it, so one person can leave behind a functioning thing instead of a deck. But the thinking has changed shape too. The human-centred design thinking that IDEO made famous is becoming something that designs with machines as readily as for people. The partner in the room is no longer only human.
And the tools are moving in the same direction. Until very recently you worked with AI alone, one turn at a time, typing into a box. That's already changing: the newest systems listen, watch, and respond in real time, across any medium. What's coming is the shared version, where a whole team and the AI work in the same live session at once. Push that far enough and a tantalising possibility appears: not just the design-engineer doing the work of an old product team, but a forward deployed product team. A small embedded crew and their machines, moving as one.
For all the novelty, the instinct runs deep, straight into the foundational ideas that shaped Made by Many's practice: Kent Beck's Extreme Programming, Steve Blank's customer development, Eric Ries's lean startup. Different vocabularies, one conviction: get the makers close to the people they're building for, and build to learn. Beck's version put the customer on site. The forward deployed model puts the engineer on site. We would say it puts the maker on site. This is not a lineage we are reaching for. It's the principle we've worked by for the better part of two decades. We simply lacked the phrase.
This is not a one-off. We have squads embedded like this right now, across several engagements, and the same thing keeps happening. One team is inside a venture-backed global business, building new experience concepts straight from the client's own data. They had spent the better part of a year trying to get there with enormous Figma boards. We got further in a few days. Another is inside a global design and engineering firm. Another at a large telecoms business. Another inside a complex global institution, working directly for its leadership.
And it goes the same way every time. The technology raised the floor on the expected things, the data wrangling and the quick prototype, and the value turned up somewhere else: in the experience layer that raw capability can't produce on its own. Nobody had specified it in advance. You could only find it by being in the room, building, and watching what became possible. That is design work, and it is the part no model vendor can sell you.
A word on "embedded," because the industry has cheapened it. It does not have to mean a body on-site; in 2026 it rarely does. What it means for us is dedicated. Most of the industry runs on a fractional model, people split across several accounts, attention sliced thin. We have never believed you can do real innovation part-time. So we dedicate small teams of very senior people. It’s closer in shape to the founding team of a high-performing startup than to an agency pod; cross-disciplinary, wearing several hats each, and trusted to make the call and own the outcome without a committee to appease. That is the real point of all of this: collapse the distance between a problem and the person who can solve it, and strip out the process tax in between. The physical shape barely matters: embedded with the client, or the client embedded with us. What matters is that the people living the problem sit inside the team, not on the far side of a handoff.
A line from David Brooks , in a recent Atlantic essay, has stayed with me: "My core belief about this whole age is that artificial intelligence will reveal what it means to be human by disclosing what AI can't do. That is the reassuring part, and the reason none of this frightens us. The better these tools become, the more sharply they outline the work that stays ours: judgement, taste, the decision about what is worth making, the human read of a room.
None of this is foreign to us. It’s not a novelty or a reinvention. It’s the thing we have always done - embed, make, stay close to the problem - brought bang up to date for a rapidly emerging new world of possibilities. We have been sceptical about plenty of things dressed up as the future. We are not sceptical about this one. We are going all in.
More soon, including why this matters most in private equity.
