Briefing / 2026-09-04 / 7 min read

Independent AI Vendor Evaluation: Why Almost No Adviser Is Independent Anymore

McKinsey, Bain, PwC and Capgemini now have formal partnerships with OpenAI or Anthropic. Who can still give an honest answer to 'should we buy this'.

Black and white geometric architecture, Independent AI Vendor Evaluation: Why Almost No Adviser Is Independent Anymore

An independent AI vendor evaluation is an assessment of whether and how a company should adopt an AI product, conducted by an adviser with no commercial relationship to any vendor or model provider. Since May 2026, when Anthropic and OpenAI launched their own consulting arms with the major consulting firms as partners, very few large advisers meet that definition.

What changed in May 2026

Within one week in May 2026, the two leading AI model companies became consulting firms. Anthropic launched an enterprise services company aimed explicitly at mid-market businesses. OpenAI launched a Deployment Company with more than $4 billion in backing, built around forward-deployed engineers and the acquisition of a 150-person consultancy. The partners in OpenAI's venture include Bain and Company, Capgemini and McKinsey. PwC has certified 30,000 staff on Claude. Accenture, BCG and others hold similar alliances.

None of this makes their advice wrong. It does mean that when a large firm tells you to deploy an AI product, it has a commercial reason to say so, and the model vendor behind the product may be its partner or its client.

Why mid-market companies feel this most

Large enterprises have procurement teams, internal architects and the budget to run competitive pilots. A company with $50 million to $2 billion in revenue usually has one or two people making the call, a vendor with a persuasive demo, and an adviser who is allied with the vendor. The Anthropic services company is explicitly targeting this segment. The pressure to buy is going to increase, and the supply of disinterested advice is going to shrink.

What an independent evaluation actually asks

An independent evaluation does not start with the product. It starts with what your customers, employees and competitors are actually doing, gathered from verified conversations with them. It then asks four questions that a vendor-aligned adviser has little incentive to press:

  1. Is the problem this tool solves a problem we have, in the size the vendor claims?
  2. What are companies like us, who bought this twelve months ago, saying now? Not case studies. Named practitioners.
  3. What breaks if the vendor's pricing, model or terms change in eighteen months?
  4. What would we do instead, including nothing?

How to tell if your adviser is independent

Ask three questions in writing. Do you have any reseller, alliance, referral or investment relationship with this vendor or its model provider? Have you disclosed every such relationship for every vendor we are considering? Will you put your recommendation in writing, signed by a named partner, including the option of not buying? An independent adviser answers all three in a paragraph.

Theory's position

Theory Intelligence has no reseller, alliance, referral or investment relationship with any technology vendor or AI lab, and The Theory Standard commits us to disclose any conflict before accepting work. We use AI in our own delivery and say exactly where. When we tell you not to buy something, nobody pays us to say otherwise. An AI vendor decision is a common use of the Decision Study.

Sources

Cite thisTheory Intelligence, "Independent AI Vendor Evaluation: Why Almost No Adviser Is Independent Anymore", theoryintelligence.com, 2026-09-04. https://theoryintelligence.com/briefing/independent-ai-vendor-evaluation/
FAQ

Questions this article answers

Which consulting firms are partnered with OpenAI or Anthropic?

Bain and Company, Capgemini and McKinsey are partners in OpenAI's Deployment Company. PwC has certified 30,000 staff on Anthropic's Claude. Accenture and BCG hold alliances with OpenAI.

What does an independent AI vendor evaluation cost?

At Theory Intelligence, a vendor decision is typically scoped as a Decision Study at $25,000 to $75,000, or as a Signal Brief at $7,500 to $12,000 for a single narrow question.

Have a decision this applies to?

A Signal Brief answers one sharp question with verified evidence in ten business days, for a fixed fee.

Start a brief