The ROI Reckoning and the 2,000 People Who Can Fix It
Something has shifted quite dramatically in the conversation around enterprise AI over the past fortnight. We have moved from tokenmaxxing to valuemaxxing. The question in boardrooms is no longer ‘which model are we using?’ but ‘where is the return on the hundreds of millions we have spent?’
The answer is exposing a critical talent bottleneck.
According to new research from executive search firm Christian & Timbers, published exclusively in TechCrunch on 30 July, there are only about 2,000 engineers in the US with the specific combination of sector knowledge, gravitas, and hands-on applied AI experience needed to consistently help enterprises see a return on their AI expenditures. Not 2,000 available. 2,000 in total.
These are the forward-deployed engineers, or FDEs. They do NOT build models in a lab. They embed directly within client organisations to build, implement, and deploy AI systems tailored to specific workflows.
The role sits somewhere between software engineering, consulting, and product deployment. And demand for them is exploding.
At the start of 2026, only 5% to 10% of companies were planning to hire FDEs. By the end of the second quarter, that number had jumped to 70%. The largest consulting and services firms are reporting a need to increase their FDE headcount by ten times, building full teams of 20 to 100 employees.
Amazon confirmed the scale of this shift the following day, announcing a $1 billion investment to build AWS Forward Deployed Engineering. This new team will work directly inside customer organisations to build and launch agentic AI systems in days rather than months. Early customers already include the NBA, the NFL, Cox Automotive, and Southwest Airlines.
The urgency behind all of this is a looming Wall Street reckoning. Jeff Christian, founder of Christian & Timbers, put it plainly: ‘This fall, Wall Street is about to say: we have given you two years to figure this out and you have not. There is no ROI. We are going to start punishing those that have spent hundreds of millions and are not generating ROI, and rewarding those that have.’
There is a further complication.
Companies are increasingly reluctant to use external FDEs from the major AI firms, fearing they will surrender their proprietary business processes to the likes of OpenAI and Anthropic. As Christian noted: ‘Everybody is concerned that if they give up their proprietary business processes, the AI firms can compete with them. So having this muscle internally is so important.’
The strategic imperative is now to build this capability in-house.
If you are a C-suite leader, your talent acquisition strategy needs to pivot immediately from hiring generic AI talent to securing the rare individuals who can orchestrate AI deployment within your specific operational context. That is a very different hire. And the window to make it is narrowing fast.
The Forward-Deployed Model Comes to Research
The scarcity of talent capable of delivering AI ROI is not limited to software engineering. The exact same dynamic is playing out in the research and insights industry, but with a twist: the forward-deployed capability is often an AI-native platform, and the scarce talent is the senior strategist who can direct it.
According to Qualtrics’ 2026 Market Research Trends report, based on responses from 3,000 researchers across 17 countries, 95% now use AI. The real difference is how confidently and strategically they are using it. General-purpose tools are starting to lose ground, with usage falling from 75% to 67% in a single year, while more specialist research platforms become part of the day-to-day workflow
This is where the forward-deployed model meets research.
As a16z recently analysed, AI-native research companies including Simile, Aaru, Quantilope, and Yabble are being deployed inside client businesses to replace traditional research functions entirely. Instead of recruiting human panels, they simulate entire societies of generative AI agents that can be queried, observed, and experimented with. Aaru already has a formal partnership with Accenture to deploy this capability inside enterprise clients.
This flattens the traditional agency pyramid. Junior execution layers are being replaced by AI. Senior strategists are directing AI agents rather than managing people. As one agency leader put it this week in Forbes:
‘We stopped hiring for the middle. AI wiped out that middle layer. Now it is a barbell, with senior thinkers on one end and sharp juniors directing AI on the other.’
The research equivalent of the FDE is a senior insights leader who knows how to question AI-generated research, govern it properly and turn it into something strategically useful at scale.
People with that mix of skills are already hard to find. As more AI-native platforms make their way into client organisations, demand for them will only grow. The firms that start identifying and holding onto this talent now will be in a much stronger position. Leave it too late, and they may find themselves scrambling for the same small pool of people.
This piece first appeared in Talent Pools Nexus, our weekly read for leaders in research, insights and consulting. Subscribe on LinkedIn.
