Most studios and agencies now need AI capability in the pipeline, and the budget for it exists. The hiring is where it comes apart. MIT’s NANDA initiative found that 95% of enterprise generative AI pilots delivered no measurable impact on profit in 2025 (MIT NANDA, August 2025). The technology rarely fails. The hiring around it does. In the searches we run for creative production companies, the same tells repeat across the briefs we see, and the strongest talent reads them long before a contract is signed.
The pattern starts with pressure. A board wants AI, a vague mandate lands on a creative director or a head of production, and an advert goes out asking for five people in one. Here’s what the best talent reads in it, and what the studios getting it right do instead.
Key Takeaways
- 95% of enterprise generative AI pilots showed no measurable profit impact in 2025; in creative production the gap sits in how studios hire and integrate, rather than in the tools (MIT NANDA)
- Companies abandoning most of their AI initiatives jumped from 17% to 42% in a single year as the early rush met reality (S&P Global)
- The strongest creative and technical AI talent now screens employers, and a bloated brief or a brain-picking interview costs the hire
- The studios getting it right hire process expertise with AI knowledge and give it time to land, rather than betting on AI expertise alone
Five roles in one advert
We’ve all seen the adverts. One AI position carrying four or five jobs, a generalist brief, and a package that doesn’t match the ask. It reads as “we need someone who does everything, and we haven’t worked out what that is”. The demand is real. PwC’s analysis of close to a billion job ads found the skills employers want in AI-exposed roles are changing 66% faster than the year before, and those roles now carry a 56% wage premium (PwC, June 2025). When the spec is mutating that fast, studios bundle every hope into one hire rather than define the role. The strongest people read that as a business that hasn’t decided what it needs.
When interviews become free consultancy
The next mistake costs people their time and studios their reputation. We’re seeing interviews run as brain-picking exercises rather than hiring processes. One senior AI hire told us about a process with a major employer, two rounds, detailed questions on tools, workflows and what rival teams were doing, then silence. We asked whether their brain had been picked throughout. It had. An iHire survey of 1,024 job seekers found 53% had been ghosted by an employer in the past year, 11% of them after multiple interview rounds (iHire, October 2025). For the people a studio actually wants, the calculation is simple. An hour spent explaining how they would build an AI pipeline for a company that never hires is an hour of unpaid consultancy. Word travels in a market this small, and the people worth hiring stop taking the call.
Short contracts for structural problems
When studios do hire, too many offer 6 to 12 week contracts for what is fundamental change to how the work gets made. We hear the same from the talent, fractional time, quick pilots rather than proper integration. As one consultant put it, “it is very new, so people need to be educated first. Companies need to taste the tech.” The instinct to start small is sound. The mistake is treating structural change as a short project. S&P Global found the share of companies abandoning most of their AI initiatives rose from 17% to 42% in a year, with the average organisation scrapping 46% of proof-of-concept projects before production (S&P Global, October 2025). A senior hire reads a six-week brief for structural work and hears a studio that hasn’t committed, and the ones worth hiring wait for one that has.
The model was never the problem
The failure is rarely the model. MIT’s NANDA researchers traced the 95% failure rate to a learning gap, the inability to fold AI into workflows, structures and culture, and noted it was “not driven by model quality” (MIT NANDA, August 2025). McKinsey found only 39% of organisations could link any profit impact to AI use, and for most that impact sat below 5% (McKinsey, March 2025). It’s the same argument we made in the AI blame game, where AI takes the blame for losses that are really about budget and structure. The technology works. The organisation around it isn’t ready.
Klarna has started hiring people back
Outside creative production the correction is already visible, and it’s worth watching. Klarna built an OpenAI assistant in 2024 that did the work of around 700 agents and let headcount fall from roughly 5,000 toward 3,000. By May 2025 CEO Sebastian Siemiatkowski told Bloomberg that cost had become too dominant a factor and quality had suffered, and the company started recruiting people back for customer service (Bloomberg, May 2025). His framing since has been about balance rather than retreat. “We can use AI to automatically take away boring jobs, things that are manual work, but we are also going to promise our customers to have a human connection” (TechCrunch, June 2025). Duolingo made the same turn, an “AI-first” memo in April 2025 became, by August, a clarification that AI augments people rather than replacing them (TechCrunch, August 2025). Creative production is watching the same lesson arrive. The teams that leaned hardest on automation are the ones now rebuilding the human judgement they cut. The value sat in the people.
Process expertise, not AI expertise
Here’s the correction in hiring terms. Studios are hiring for “AI expertise” when they need process expertise with AI knowledge. Gartner found that through 2026 organisations will abandon 60% of AI projects that aren’t supported by AI-ready data, and that 63% either lack the right data practices or are unsure they have them (Gartner, February 2025). Without context, the tools guess. As the same consultant told us, “without any context or preparation, they ask, and then they get an output, and the AI is just guessing.” The person who fixes that understands how the work gets made first and the model second.
The judgement comes before the commitment
The people we’ve placed into these roles got in through other skills, production, post, development, a real grasp of the craft, then proved the value of AI workflows once inside. The studios that got a taste extended those contracts and invested properly. That’s the order that works. Hire the judgement, prove the value, then scale the commitment. We spend most of our time matching that kind of senior and C-level talent to creative production companies that have moved past the rush and started defining the work, often through confidential search. The roles worth hiring combine creative and production judgement with AI fluency, and what we deliver is built around finding exactly that. The strongest people can tell these studios apart in the first conversation, and they choose them. Stop treating AI as a magic hire. Start treating it as capability that needs the right person and the room to prove it.
