AI proposes, people decide: accountability in AI-assisted consulting

AI makes analysis faster and broader than any team could manage alone. It also raises a question that every client should ask: when an AI-assisted recommendation turns out to be wrong, who is responsible? Our answer is fixed and written into how we work.

The AI proposes and correlates. A senior consultant checks every finding, signs the advice and stays accountable for it.

What the AI is for

In our engagements, AI does the work that benefits from speed and breadth: reading large volumes of documents and data, finding patterns across departments, linking symptoms to candidate causes, and drafting summaries with their sources attached. It is very good at proposing. It is not trusted to decide.

What people are for

• Testing each candidate finding against the evidence and against the people who do the work
• Rejecting findings that are weakly supported, however plausible they sound
• Weighing context the evidence does not contain: politics, history, priorities
• Signing the final recommendation and answering for it

Access follows responsibility

Accountability is also built into our tools. In both UniversAIDS and UniversAIMS, what a consultant can do follows their role. Senior consultants run evidence synthesis; field consultants work on the departments assigned to them; only a partner can open an engagement's structure or close it and purge its raw evidence.

Why clients should insist on this

A recommendation nobody signs is a recommendation nobody owns. When a named person is accountable, the analysis is checked harder, the reasoning is explained better, and the client always knows who to call. That is worth more than any amount of speed.