A research firm
with a name on it

Theory Intelligence was founded in 2019 on a plain observation: the research and advisory market had become very good at producing documents and much worse at producing evidence anyone could check. Seven years and one generative AI wave later, that observation is the whole industry's problem.

When we started, the gap was between what a client paid for and what had actually been verified. Panels were already filling with low-quality respondents, expert network profiles were already self-reported, and the footnotes in most strategy decks already led to other strategy decks. We built our method around a fixed rule: nothing goes in a deliverable unless a named person has traced it to a source they have seen. That rule did not change when generative AI arrived. It became the reason clients called.

In 2025 Deloitte Australia refunded part of a A$440,000 government fee after a report cited a court judgment that did not exist. In 2026 PwC published reports whose footnotes led nowhere. In August 2026 Source Global Research asked 3,887 companies whether they would trust a report prepared with AI, and 70 percent said no. At the same time, Kantar reported that 30 to 40 percent of online survey responses are compromised, and academic tests found that bots pass survey quality checks 99.8 percent of the time.

None of this means AI is bad at research. It means the profession stopped verifying. Theory exists to do the verifying, publicly and on the record, and to put a person's name on the conclusion.

What we believe

Evidence is a chain, not a document. A claim is only as good as the source behind it and the person who checked that source. Judgment is a signature, not a slide. A recommendation with no one accountable for it is a summary, and clients have enough of those. Independence has to be structural. We take no fees, referrals or partnerships from technology vendors or AI labs, so when we tell you a tool is not worth buying, nothing about our business changes.

How we are built

Theory is a small firm by design. Every engagement is led by a named researcher who gathers the evidence, does the analysis and signs the recommendation. We use AI for the parts of the work where a machine is reliably better than a person: scheduling, transcription, coding of notes, first drafts of routine sections. We disclose that use in every deliverable. We do not use AI as a source of facts, and we do not use synthetic respondents in place of people.

Because our method is fixed, our prices can be fixed too. A Signal Brief costs USD 7,500 to 12,000. A Decision Study costs USD 25,000 to 75,000. An Intelligence Retainer costs USD 6,000 to 15,000 a month. The scope that goes with each price is published on this site.

Who we work with

Investors, executives and operators facing a decision where being wrong is expensive: an acquisition, a market entry, a pricing change, a vendor selection, a product bet. Most of our work is in private equity, B2B software, insurance and healthcare, but the method transfers to any sector where the people who know are not the people who publish.

What we are not

We are not an expert network. We do not sell you calls and leave you to make sense of them. We are not a strategy consultancy with a hundred slides and a bench to keep busy. We are not a synthetic data vendor. We are a research firm that produces verified evidence and signed judgment, at a published price, in a fixed time.