AI Collaboration Architect

This is a projected career: someone who redesigns how a company's people and AI systems share work, deciding which tasks go to AI, where a person checks the output, and who steps in when something goes wrong. There is no government job code for it and no hiring data yet.

The demand case is real. Deloitte's 2026 technology predictions say that over the next 12 to 18 months more businesses will scale up systems of cooperating AI agents while 'keeping humans in the loop', and will need to define 'new human skills and responsibilities for agent training, orchestration, oversight, and governance'. It expects companies to sort work along a range from humans 'in the loop' to 'out of the loop', which is exactly the kind of decision this job makes. Two cautions: Deloitte sells AI consulting, so it benefits from this view, and it also cites an estimate that more than 40% of today's AI agent projects could be cancelled by 2027 over cost and complexity.

The risk case is that much of the day-to-day work, such as mapping workflows, analyzing processes and writing recommendations, is the same work management consultants do, and that is among the most AI-exposed work in the economy. Microsoft's 2025 study ranks management analysts, the government's job code for consultants, 23rd of 785 jobs for how much of the work AI chatbots can help with. Anthropic's March 2026 study, which tracks what people actually use its Claude AI for at work, puts them at about 24% of tasks (94th of 756). Both studies come from AI companies. The Bureau of Labor Statistics still projects management analyst jobs to grow 10% from 2025 to 2035, much faster than average.

We move this from 2 to 3 (Moderate). Our old reasoning already said the analytical core of the job was automatable but scored it as if it were not. AI adoption creates demand for this work, and the persuading and politics of changing how a company works stay human, but the analysis a junior person would start on is highly exposed. We score management consultants a 4 and AI agent orchestration specialists, a more technical AI-created job, a 2; this sits between them. What would change the score: companies creating and growing dedicated human-AI workflow teams would move it to 2. The role being folded into ordinary management or consulting as AI becomes routine would move it to 4.