The platforms often returned entirely different shortlists
OpenAI, Gemini and Perplexity had no recommended firm in common 60% of the time.
How AI chooses UK financial advisers
AI is already shaping which advisers enter a prospective client’s shortlist. Studio Baggio tested how ChatGPT, Gemini and Perplexity choose UK financial advisers across 50 high-intent prospective-client questions.
These covered investing £500,000+, inheritance tax, retirement income and planning before and after a business sale.
Four headline findings
Advice firms supplied much of the expertise. Directories, publishers and commercial rankings often carried the visibility and shaped the shortlist.
The credit gap
In 97% of guidance answers that cited a panel firm’s website, AI used the firm’s expertise without naming the firm to the buyer.
Panel-firm websites were cited 76 times across 117 guidance answers. The firm was named in only two of those 76 citations. When AI selected a firm directly, only 29.7% of selections cited that firm’s own website. The remaining selections relied on other sources, including directories, rankings and reviews.
What happened when AI was asked to choose
The firms changed with the platform, the buyer’s question and the repeated run. Some answers did not name a firm at all.
OpenAI, Gemini and Perplexity had no recommended firm in common 60% of the time.
When AI was asked to choose a firm nationally, 46 of 180 answers returned guidance, directories or authorities instead of a financial advice firm.
Only four of 303 candidates appeared across all four national buyer-need categories. Most visibility was narrow or one-off.
Who shaped the answer
Directories, public bodies and commercial pages repeatedly supplied the information AI used to explain decisions and build firm shortlists.
For example, Nephos Group’s “10 best financial advisers in the UK” page was cited 21 times and tied as the most-cited individual direct page in the firm-selection questions. The commercial ranking placed Nephos first and showed no visible selection methodology.
168 of 450 answers
127 of 450 answers
118 of 450 answers
66 of 450 answers
47 of 450 answers
This finding concerns source influence, not adviser quality. Commercial pages from one financial-services group sat alongside Unbiased, MoneyHelper, VouchedFor and the FCA while most of the established 150-firm panel remained invisible.
Live follow-up experiment
On 7 August 2026, Studio Baggio began a live follow-up test using three Calm Authority comparison guides. Each discloses its selection method, links the evidence behind every firm and accepts no paid inclusion or placement. The test asks whether pages built that way can enter the sources cited by ChatGPT, Gemini and Perplexity.
Early results show that all three Calm Authority guides are now appearing in AI search. Where Calm Authority appeared alongside the commercial ranking pages discussed above, it ranked higher seven out of eight times. We will continue to track how that changes.
Different needs, different leaders
The firm recommended most often changed with the buyer’s financial need.
General financial advice
The Private OfficeMost often recommended · 14 of 45 AI answersWealth and investing
Evelyn PartnersMost often recommended · 21 of 45 AI answersPensions and retirement
Hargreaves Lansdown Financial AdviceMost often recommended · 6 of 45 AI answersLife events and specialist planning
Evelyn PartnersMost often recommended · 11 of 45 AI answersWhy this matters now
Buyers are increasingly using AI to research options, compare firms and decide who to choose. The route to the shortlist is shifting, and being part of the consideration set is becoming increasingly commercially valuable.
of B2B buyers using AI say it has shaped their vendor shortlist, while 83% say it has influenced their final vendor decision.
Semrush, 2026of consumers have replaced traditional search engines with generative AI tools as their go-to for product and service recommendations, up from 25% in 2023.
Capgemini Research Institute, 2025of Google searches are expected to include AI-generated summaries by 2028.
McKinsey, 2025AI-referred visitors converted better than non-AI traffic in July 2026.
Adobe Digital Insights, U.S. retail dataThe full question set
The study followed the buyer journey from understanding a financial need to finding and choosing a firm. Open any group to see the exact wording.
Questions about costs, pensions, inheritance, divorce and other financial decisions.
Questions asking AI to identify firms by need, specialism or location.
Direct tests of which firms AI recommended for a specific buyer need.
Selection breadth
Search for a firm, filter by advice area or a specific buyer need.
Showing 20 of 303 firms and advisers found in national answers.
| Overall rank | Candidate | Advice areas reached | General | Wealth | Pensions | Life events |
|---|---|---|---|---|---|---|
| 1 | 4 of 4 | 1% | 7% | 0.2% | 3.2% | |
| 2 | 4 of 4 | 4.5% | 2.3% | 1.1% | 2.2% | |
| 3 | 3 of 4 | 2.5% | 4.3% | 0% | 2.1% | |
| 4 | 4 of 4 | 2.4% | 0.6% | 3.7% | 0.8% | |
| 5 | 2 of 4 | 0.8% | 5.9% | 0% | 0% | |
| 6 | 2 of 4 | 0% | 0% | 3% | 2.6% | |
| 7 | 3 of 4 | 0.9% | 3.1% | 0% | 1.3% | |
| 8 | 2 of 4 | 1.5% | 0% | 0% | 3.7% | |
| 9 | 3 of 4 | 2.6% | 2.3% | 0% | 0.2% | |
| 10 | 3 of 4 | 0.2% | 3.6% | 0% | 1.1% | |
| 11 | 3 of 4 | 1.2% | 1.9% | 0% | 1.4% | |
| 12 | 1 of 4 | 0% | 0% | 0% | 4.4% | |
| 13 | 3 of 4 | 0.4% | 3.3% | 0% | 0.6% | |
| 14 | 4 of 4 | 1.1% | 1.4% | 0.7% | 0.7% | |
| 15 | 3 of 4 | 0.7% | 1.6% | 0% | 1.4% | |
| 16 | 2 of 4 | 0% | 0% | 3% | 0.2% | |
| 17 | 2 of 4 | 0% | 0% | 2.6% | 0.4% | |
| 18 | 2 of 4 | 2.4% | 0.4% | 0% | 0% | |
| 19 | 1 of 4 | 2.6% | 0% | 0% | 0% | |
| 20 | 1 of 4 | 0% | 0% | 2.6% | 0% |
What firms should take from this
The firms in this study already possess expertise, qualifications and market authority. AI can only use evidence it can retrieve, understand and connect to a buyer’s question.
Without both, a firm may help produce the answer while remaining invisible to the buyer.
SEO and AI Search Opportunity Audit
Studio Baggio audits the buyer questions that matter, identifies the sources shaping the answers and sets out the evidence your firm needs to enter those consideration sets.
Discuss your firm's visibility