top of page

GEO study: which CRM do AI assistants recommend to small businesses in the UK?

Oct 8
18 min read

Five AI engines, 37 CRM brands and 21 vendor websites: the brands that answers to UK small-business CRM questions name, the sources behind them and the brands the answers never reach.


AI, TELL ME! CRM Visibility Study, United Kingdom. Written by Natali Smith, co-founder of AI, TELL ME! Published 8 October 2026. Answer data collected 1 October 2026, website data collected 7 October 2026.


Short answer. Across 225 AI answers to CRM questions of the kind a UK small business asks, this GEO study found that five AI engines named 37 CRM brands 2,087 times, and Pipedrive and HubSpot together hold 49.2% of those mentions. Below those two, the brands named vary by engine and by the kind of CRM the buyer asks for. In this small sample no website property measured showed a convincing link to how often a brand is named, and the three CRM vendors named in no answer all had reachable websites.

Cover: "Which CRM systems are recommended by AI assistants?" with three figures from the study (225 AI answers, 37 brands named and 49.2% of brand mentions going to two brands), beside a search box asking "Best CRM for a small team?" whose four source cards (website, document, quote and video) feed one glowing point, above a laptop showing a CRM app on a desk in front of a London skyline at dusk.

A small-business owner in the UK opens an AI assistant and asks which CRM to buy. Across 45 questions of that kind, put to five AI engines on 1 October 2026, 37 CRM brands were named 2,087 times between them. Pipedrive and HubSpot take the top two places by a wide margin, with Zoho CRM third. Three further CRM vendors selling in the UK were named in none of the 225 answers. The answers doing the naming cited 295 distinct websites, and 32 of those belong to the CRM brands themselves.


Both halves of this study ran on our own tools

AI, TELL ME! is a full-cycle AEO (answer engine optimisation) and GEO (generative engine optimisation) agency, and we operate our own monitoring and website assessment tools, which produced both halves of this study: the monitoring platform that records what each AI engine says and cites, and the AI Access Report that records what a probe using the engines’ crawler names can reach and read on a website.


We set the methodology, so every figure below traces back to the run that produced it, and the prompt set, the sampling rule, the evidence grades and the size of the run are documented in the methodology section at the end.


The monitoring ran in English with a UK geo target:


  • Three segments: CRM for SMBs, Sales CRM for SMBs and CRM with automation for SMBs.

  • Five engines: ChatGPT, Google AI Mode, Claude, Gemini and Google AI Overview.

  • 225 answers to buyer questions such as which CRM suits a small team, what to compare for lead tracking and which tool automates follow-ups.


Twenty-one CRM vendor websites were measured alongside the answers, 18 of them belonging to brands the answers name and 3 to brands the answers never name.


Definitions: how to read the numbers in this study

Eight measures carry the findings below.


  • A mention is one count of a brand in an answer, and one answer can carry several mentions of the same brand.

  • Share of voice is the share of all mentions that belongs to one brand. It describes conversational presence and says nothing about traffic or revenue.

  • Answers naming a brand counts the answers that name the brand at least once.

  • A domain citation is one cited website domain in one answer, counted once however many of its pages the answer links.

  • Source group sorts every cited domain as an own site, another company’s site or an independent source (review, press, community and every other domain).

  • Content extraction score is how much of a page a retrieval pipeline recovers as content once menus, filters and footers are stripped, as a score out of 100. It differs from the percentage of visible text the extractor keeps, which is reported separately.

  • Passage quality is how well a page survives being cut into the passages an answer is built from.

  • Structured data score is the report’s score for structure and markup: JSON-LD, machine-readable publication dates, one H1 per page, unbroken heading levels, a declared language and a page title and description.


Pipedrive and HubSpot take nearly half of all CRM brand mentions in the UK

Bar chart of share of voice for 40 CRM brands across 2,087 mentions on five AI engines: Pipedrive 24.8%, HubSpot 24.4%, Zoho CRM 12.1%, Freshsales 7.1% and Salesforce 6.0%. Workbooks, SugarCRM and SuiteCRM sell in the UK and were named in no answer.
Figure 1. Share of voice across all five engines, 2,087 mentions. Brands below the line sell in the UK and were named in no answer.

Pipedrive holds 24.8% of the 2,087 mentions and HubSpot 24.4%, which puts the two at 49.2% between them. At seven mentions apart, this study reads them as level (the counting basis is set out in the methodology section).


The next eight brands hold 39.9% between them, which brings the top ten to 89.1% of all mentions, and 23 of the 37 named brands sit below 1% each. Three CRM vendors at the foot of the chart (Workbooks, SugarCRM and SuiteCRM) sell in the UK and were named in none of the 225 answers.


Table 1. The ten most-mentioned CRM brands, all five engines.

Brand

Mentions

Share of voice

Answers naming the brand (of 225)

Engines naming the brand (of 5)

Pipedrive

517

24.8%

208

5

HubSpot

510

24.4%

194

5

Zoho CRM

252

12.1%

160

5

Freshsales

149

7.1%

81

5

Salesforce

125

6.0%

65

5

Bigin

91

4.4%

49

5

Capsule CRM

69

3.3%

40

5

monday CRM

59

2.8%

40

5

ActiveCampaign

48

2.3%

23

4

Close CRM

40

1.9%

14

5


Zoho CRM is the usual third name

Zoho CRM holds 12.1% of mentions, ahead of Freshsales in fourth place with 7.1%, and it is the third name on four of the five engines (Salesforce is third on Claude). Bigin is a Zoho product, and the two Zoho lines together hold 16.4% of mentions.


Below the top ten the field is thin

Less Annoying CRM, Attio, Nutshell and Salesflare hold between 21 and 31 mentions each, and the 23 brands under 1% hold between 1 and 16 mentions each. Three brands were named once (crm2, EngageBay and totalCRM), and a single mention in 225 answers records one event and describes no position.


The segment changes who appears below the leaders

Table 2. Answers naming each brand, by segment (75 answers per segment).

Brand

CRM for SMBs

Sales CRM for SMBs

CRM with automation for SMBs

ActiveCampaign

0

2

21

Capsule CRM

24

8

8

Less Annoying CRM

16

4

3

Close CRM

1

11

2


What a buyer asks changes which brands appear below the two leaders (Table 2). ActiveCampaign is named in 21 of 75 answers in the automation segment and in none of 75 in the general segment, and Capsule CRM, Less Annoying CRM and Close CRM each concentrate in one segment as well.


A chi-square test on the ten brands named in the most answers, corrected for the number of tests, passes ActiveCampaign, Capsule CRM and Less Annoying CRM, and Close CRM (outside those ten, tested after the fact with Fisher’s exact test) gives a result of similar strength. Pipedrive, Bigin, HubSpot, Zoho CRM, Freshsales, Salesforce and monday CRM show no segment difference that survives the correction (the test statistics are in the methodology section). These are presence rates within a segment and carry no ranking.


The five engines agree on the top two and differ below them

Heatmap of each engine's brand mentions by brand: Pipedrive and HubSpot together hold 39.3% on Claude and 56.8% on ChatGPT, Salesforce holds 12.0% of Claude's mentions and 1.5% on each Google engine, and Bigin holds 0.3% of ChatGPT's mentions and 8.6% of Google AI Mode's.
Figure 2. Share of each engine’s brand mentions per brand, in overall order. Shade runs from white (0%) to black (30%) and a zero is printed as 0.0.

Pipedrive and HubSpot are the two most-mentioned brands on all five engines, and their combined share runs from 39.3% on Claude to 56.8% on ChatGPT (Table 3). Below them the engines part ways: Salesforce holds 12.0% of Claude’s mentions and 1.5% of the mentions on each of Google AI Mode and Google AI Overview, while Bigin holds 0.3% of ChatGPT’s mentions and 8.6% of AI Mode’s. A brand that checks only one engine sees one fifth of the answers in this study.


Table 3. Engine profile: brands named, sources cited and the share held by the two leaders.

Engine

Mentions

Brands named

Pipedrive and HubSpot combined

Distinct domains cited

Domain citations

ChatGPT

625

15

56.8%

37

163

Gemini

477

21

53.2%

63

149

Claude

384

31

39.3%

129

320

Google AI Mode

338

20

41.1%

32

103

Google AI Overview

263

18

48.7%

126

307

All engines

2,087

37

49.2%

295

1,042


Breadth and depth differ by engine

Claude names the most brands and cites the most distinct domains, and ChatGPT names the fewest brands (Table 3). Google AI Overview cites 126 distinct domains while naming 18 brands, the highest ratio of sources to brands of the five, and every engine returned source links.


Eleven brands appear on Claude alone, and six of them have .uk domains

Of the eleven brands that only Claude names, six (Gather360, Orbis CRM, popcorn CRM, YourPCM, crm2 and totalCRM) have a .uk domain, and those six domains carry 23 citations, all on Claude. The eleven brands share 2.2% of all mentions, so this is a pattern in the tail and says nothing about the leading brands.


Company websites supply two-thirds of the sources, and the mix depends on the engine

Stacked bars of domain citations by source group: own sites supply 55.8% of ChatGPT's citations and 53.8% of Claude's, against 20.8% for Gemini, 10.1% for Google AI Overview and 7.8% for Google AI Mode, and across all engines own sites supply 32.0%, other companies 36.0% and independent sources 32.1%.
Figure 3. Share of domain citations by source group, per engine. Own sites plus other companies’ sites make up the company-site total.

Citations split into own sites, other companies and independent sources

The 1,042 domain citations fall on 295 domains, and company websites supply two-thirds of them (67.9%). They split into the own sites of tracked CRM brands (32.0%) and other companies’ sites (36.0%), and independent sources supply the remaining 32.1%. Counted by source weight (the monitoring platform’s own percentage for each cited source), own sites hold 29.7% and independent sources 35.7%. The methodology section sets out the grouping rule behind the split.


Own sites supply more than half of ChatGPT’s domain citations (55.8%) and of Claude’s (53.8%), against 7.8% of Google AI Mode’s (Figure 3). ChatGPT cites 37 distinct domains in all, so its high own-site share rests on a narrow set of sites.


Google AI Mode and AI Overview supply almost all Reddit and YouTube citations

Google AI Mode and Google AI Overview between them supply 34 of the 36 Reddit citations and all 36 YouTube citations. Gemini cites Reddit twice and YouTube never, and ChatGPT and Claude cite neither. Each of the two domains takes 3.5% of all citations.


Citations are spread across 295 domains, and press has the highest share on .uk domains

Two bar charts of 1,042 domain citations: pipedrive.com (49), hubspot.com (47) and salesforce.com (47) lead the ten most-cited domains, which hold 33.3% of citations together, and 161 of the 295 cited domains (54.6%) are cited once.
Figure 4. Domain citations: the ten most-cited domains and all 295 cited domains by number of citations. Black bars are domains of tracked brands.

Ten domains hold a third of the citations (33.3%), and 161 of the 295 domains are cited once. Six of the ten are own sites of tracked brands (pipedrive.com, hubspot.com, salesforce.com, zoho.com, capsulecrm.com and freshworks.com), and the other four are pcmag.com, youtube.com, reddit.com and startups.co.uk.


Domains ending in .uk number 57 and carry 17.7% of citations, and press has the highest .uk share: 43 of the 67 press citations (64.2%) sit on .uk domains, supplied by four UK press domains, against a smaller share of review, company and community citations. The .uk test is a suffix test, so it misses UK sites that sit on .com.


Most vendor websites served AI crawler requests and scored lowest on page structure

The 21 sites appear as Site A to Site U, with Sites A to R belonging to brands named in the answers and Sites S to U to brands named in none. The methodology section explains the labelling and the limits of a single probe.


Heatmap table of five readiness scores for 21 anonymised vendor sites, Site A to Site U, each probed once on 7 October 2026: structured data is the lowest score with a median of 66 against 100 for crawler access and for text without JavaScript, and cells with no score are labelled not measured.
Figure 5. Readiness scores for Sites A to U, one probe each on 7 October 2026. Shade runs from white (0) to black (100) and not-measured cells are labelled.

Access and text sit near the ceiling, structure is lowest

Crawler access and text without JavaScript sit near the ceiling, each with a median of 100, and structure is the lowest of the five scores, with a median of 66 (Table 4). Four sites reach 83 or more on structure, so the weakness is widespread and has exceptions. In plain terms, the sites mostly let crawlers in and hand them readable text, and the weak spot is the markup that tells a machine what a page contains.


Table 4. Score ranges across the sites measured. Sites with no score for a measure are labelled not measured in Figure 5 and are left out of every range and median.

Score

Sites measured

Range

Median

Crawler access

17

88 to 100

100

Text without JavaScript

17

92 to 100

100

Content extraction

18

55 to 100

89

Passage quality

18

82 to 100

95.5

Structured data

18

34 to 92

66

Speed (lab)

3

35 to 94

60


Ten sites served every crawler name and three refused all of them

Table 5. Crawler replies by site, as counts (21 sites, 14 fetching crawler identities each).

Reply pattern

Sites

Served all 14 identities

10

Refused one training crawler (Bytespider) and served the other 13

5

Refused five training crawlers and served all 8 citation crawlers

1

Served 9 of 14, with 5 checks that did not land

1

Served 5 named AI crawlers and refused 9

1

Refused all 14 names

3


Two of the 21 sites are withheld from any verdict about AI crawlers because they refused every request, plain ones included (the methodology section defines the evidence grades). The three sites that refused all 14 names belong to brands the answers name and whose own domains the engines cite, and one of them served the probe’s own name, which marks its refusal as an identity check. A refusal in a probe describes what a test client received from one address on one day and settles nothing about what a live crawler receives.


Page signals differ between sites

Across the 18 sites with pages read:


  • JSON-LD sits on every sampled page at 5 sites and on under a third of the pages at 6.

  • A machine-readable date sits on every sampled page at 3 sites and on none at 4, and the two signals do not travel together: two of the five sites with JSON-LD on every sampled page carry dates on only some pages or on none.

  • The extractor keeps between 21% and 89% of the visible text (median 46.5%).


Of those 18 sites, 12 publish an llms.txt file.


Three brands appear in no answer, and all three sites were reachable

The sites of Workbooks, SugarCRM and SuiteCRM were reachable in the probe: crawler access scores are 98, 100 and 100, and text without JavaScript scores 100 on all three, so access gives no reason for their absence. Two of the three have lower extraction scores than every site of a named brand, and on structure one scores below every named-brand site while two score at or above the named-brand median of 66. One of the three was read on 7 pages against 21 for most sites, so its scores rest on a thinner sample.


In this small sample, no measured site property shows a convincing link to how often a brand is named

Eleven website measures, from the readiness scores in Table 4 to page signals such as JSON-LD and machine-readable dates, were set against three outcomes (mentions, answers naming the brand and engines naming the brand), which gives 33 tests, each a Spearman rank correlation (the full list of measures is in the methodology section).


Among the sites of named brands (14 to 15 per test), none of the 33 reaches p < 0.05, and the largest coefficient is 0.51. When the three never-named sites join the table, 9 of the 33 tests reach p < 0.05, and those nine results rest on three added points with zero mentions and none survives a correction for 33 tests. With 14 to 18 sites no coefficient supports a claim: in this small sample no convincing link could be detected, and a sample of this size can neither confirm a link nor rule one out.


The eight most-mentioned brands are named on all five engines and have their own sites cited

The eight most-mentioned brands (Pipedrive, HubSpot, Zoho CRM, Freshsales, Salesforce, Bigin, Capsule CRM and monday CRM) are each named on all five engines, and each has its own domain cited, from 49 times for Pipedrive to 3 for Bigin.


Across the 37 named brands, own-domain citations track mentions (Spearman rank correlation 0.56, p = 0.0003), and the relation holds without the two leaders and across the 18 named brands whose websites were probed. The ratio varies widely: ActiveCampaign has no own domain cited among its 48 mentions, while Nutshell has 16 own-domain citations on 22 mentions.


What the results mean for a CRM brand

The results point to four checks a CRM brand can run on its own category before it commissions any optimisation:


  • Check visibility for each kind of question a buyer asks, because ActiveCampaign is named in 21 of 75 answers to automation questions and in none of 75 general ones, and a single overall figure hides that spread.

  • Compare the engines one by one, because Salesforce holds 12.0% of Claude’s mentions and 1.5% of Google AI Mode’s, and a brand that checks only one engine sees one fifth of the answers in this study.

  • Study the pages the answers cite in your own segment, because own sites supply 32.0% of all citations and the other two-thirds sit on other companies’ sites and independent sources such as review, press and community pages.

  • Repeat the measurement on a fixed set of questions, because this study is a single run and a change in share of voice only counts when the questions have stayed the same.


A website audit comes after those four checks: in this small sample no website property measured showed a convincing link to how often a brand is named, so the audit confirms that engines can reach and read the site without showing what moves share of voice.


After an audit, the work moves from measurement to a plan

A study of this kind describes a starting position, and for a single brand AI, TELL ME! runs the work that follows:


  • AEO Monitoring shows whether and where a brand appears in AI-generated answers, how it is described and which sources shape each response.

  • The GEO Report evaluates the brand’s website, content and external presence against the signals the engines use, and says what helps the brand appear and what holds it back.

  • The AEO & GEO Roadmap turns the findings into a prioritised 90-day plan covering technical structure, content, messaging and external presence.

  • Implementation carries the plan out alongside the brand’s own team, as a fully outsourced function or as a blend of both, covering technical fixes, content, external authority building and ongoing optimisation.


Monitoring then repeats on the same prompt set, so a change in share of voice or in citations can be read against the same questions.


The data do not support copying the leaders

The visible brands have their own sites cited often, which invites a brand to copy what the leaders publish. Three findings argue against treating that as a recipe:


  • Own sites supply 32.0% of all citations, so two-thirds of the sourcing sits on pages a brand does not control, and 161 of 295 domains were cited once.

  • Among the sites of named brands in this small sample, none of the 11 site measures shows a convincing link to how often a brand is named, and the three never-named sites were reachable in the probe, so matching a leader’s technical profile carries no demonstrated advantage.

  • The own-domain correlation has no known direction, and a leader’s cited site may follow from its visibility as easily as cause it.


These findings against other published research

Five published analyses bear on what this study found: three match it and two diverge.


SparkToro and Gumshoe ran 12 brand-recommendation prompts across ChatGPT, Claude and Google’s AI surfaces, each repeated 60 to 100 times per platform, and reported in January 2026 that the chance of the same list of brands appearing twice is under 1% for ChatGPT and Google’s AI (slightly higher for Claude) and the chance of the same list in the same order under 0.1%. SparkToro called visibility percentage across many prompts a reasonable metric and ranking position in AI unreliable. That matches this study’s choice to treat single-run figures as directional and to read the segment results as presence rates.


Two analyses of structured data match the absence of a convincing link in this study. Ahrefs reported in May 2026 that 1,885 pages which added JSON-LD showed no clear increase in citations against matched control pages 30 days later, on Google AI Overviews, Google AI Mode and ChatGPT (AI Overviews showed a small, statistically significant decline of 4.6%, and the changes of 2.4% on AI Mode and 2.2% on ChatGPT were indistinguishable from zero). Search Atlas reported in December 2025 that domains with complete schema coverage were cited no more often than domains with little or none, on three engines. Ahrefs studied pages that already had 100 or more AI Overview citations, and Search Atlas measured whether schema was present without grading its quality.


Yext’s analysis of 155.5 million location-grounded citations across 1,623 brands, published in July 2026 for January to March 2026, found brand websites supplying 52.0% of Gemini’s citations and 50.9% of Perplexity’s, with Anthropic’s model the lowest of the four. Here own sites supply 32.0% of citations overall, 20.8% on Gemini and 53.8% on Claude. The samples differ, with location-grounded queries about local brands there and CRM buyer questions here, which is a reason to measure a category before applying a cross-industry benchmark to it.


Peec AI’s analysis of 30 million US sources, published in March 2026, ranks Reddit first and YouTube second, and lists Reddit in the top five on all five platforms it covers and YouTube on four of them. Here Reddit and YouTube each take 36 of the 1,042 citations (3.5%), almost all from Google AI Mode and Google AI Overview, and ChatGPT cites neither (Claude was not in the Peec set).


Frequently asked questions

Do AI assistants name different CRMs depending on the question?

Yes, below the two leaders: ActiveCampaign is named in 21 of 75 answers to automation questions and in none of 75 answers to general small-business CRM questions, and Close CRM appears mainly in sales-segment answers. Pipedrive and HubSpot are named in 63 to 73 of the 75 answers in every segment.


Why does a CRM appear in one AI assistant and not in another?

The engines name different mixes of brands: Salesforce holds 12.0% of Claude’s mentions and 1.5% of Google AI Mode’s, and ChatGPT names 15 brands in all against 31 for Claude.


Do AI assistants cite a CRM vendor’s own website?

Yes, and how much depends on the engine: own sites supply 32.0% of all 1,042 domain citations, 55.8% of ChatGPT’s, 53.8% of Claude’s and 7.8% of Google AI Mode’s.


Does blocking AI crawlers keep a CRM out of AI answers?

Not in this probe: all three sites that refused all 14 crawler names belong to brands the answers name and whose own domains the engines cite. A refusal in a probe settles nothing about live crawlers, and server logs are the only place to see what they receive.


Does structured data help a CRM appear in AI answers?

In this small sample no convincing link between structured data and a brand’s mentions could be found: none of the 33 tests of site properties against how often a brand is named reaches statistical significance among the sites of named brands, and Ahrefs’ test of 1,885 pages that added JSON-LD found no clear increase in citations either.


Methodology and limitations

  • This study in numbers: one category (CRM software for small businesses) in one market (the United Kingdom), three segments, 45 prompts, five AI engines, 225 answers that between them name 37 brands, 2,087 mentions, 1,042 domain citations on 295 domains and 847 cited URLs, plus 21 vendor websites probed. Citation figures cover the 214 of 225 answers that carried sources, and three vendors the answers never named were added to the brand list at zero.

  • Fifteen prompts per segment is a test scale, chosen to establish the shape of a category and the source set behind it. Working monitoring runs 50 prompts per segment, and every segment is monitored on its own, because presence differs between segments (Table 2). The run is a single reading from 1 October 2026, and run-to-run variance is a property of these systems, so single-run figures are directional.

  • Share of voice divides each brand’s mentions by the 2,087 mentions in all answers, or by one engine’s mentions in the engine view. Answers naming a brand are counted out of 225 (75 within a segment), and a citation percentage divides by the 1,042 domain citations unless the text names an engine or a source type as its base. Brand counts follow the monitoring platform’s brand extraction, and close pairs are read as level.

  • The segment comparison is a chi-square test on the ten brands named in the most answers with a Bonferroni threshold of 0.005, which ActiveCampaign (p < 0.0001), Capsule CRM (p = 0.0004) and Less Annoying CRM (p = 0.0005) pass and Pipedrive (p = 0.044) and Bigin (p = 0.036) do not. Close CRM sits outside those ten and was tested after the fact with Fisher’s exact test, sales segment against the other two, which gives p = 0.0005. The eleven website measures in the 33 correlation tests are the five scores in Table 4 other than speed, the share of visible text the extractor keeps, four page-signal shares (JSON-LD, machine-readable date, heading levels and single H1) and the eligible URLs in the sitemap. Among the sites of named brands the smallest p is 0.053, and with the three never-named sites added (n = 17 to 18), 9 of the 33 tests reach p < 0.05 against 1.65 expected by chance and none survives a Bonferroni threshold of 0.0015 in either set. Without the two leaders the own-domain correlation is 0.48 (n = 35, p = 0.003), and across the 18 named brands whose websites were probed it is 0.67 (p = 0.002).

  • A cited domain is an own site when it maps to a tracked brand by domain name (freshworks.com, monday.com and zoho.com are parent-company sites and count as own, and one HubSpot-hosted file domain is mapped to HubSpot), another company’s site when the export types it as a company and an independent source for every other domain. Six named brands (ActiveCampaign, Brevo, Copper CRM, Keap, Odoo and EngageBay) have no own domain among the 295 cited.

  • Each of the 21 vendor sites received one AI Access Report on 7 October 2026: 14 fetching crawler identities (8 citation crawlers and 6 training crawlers) sent to the home page from one address and compared with a desktop browser’s reply, plus the home page and 20 pages drawn from the sitemap across its page templates, three of them opened in a headless browser on 17 sites and one on the other four. Fifteen sites were read on 21 pages and three on 2, 7 and 20 pages. Three refused the plain requests and have no page-based scores, and each page-signal share describes the sampled pages only. Scores are a relative scale for tracking change over time and have not been calibrated against how often a site is cited.

  • Access findings carry one of three evidence grades: site-attributable (the evidence sits in a page’s HTML or a static file, or a server answered some crawler names and refused others inside one run from one address), withheld (the site refused every request, so no verdict about AI crawlers is possible) and contradicted (the finding sits against other evidence in the data). Site-level findings are published as Site A to Site U, because one probe from one address cannot prove what a live AI crawler receives. The letters were assigned by a seeded shuffle in two blocks, and no letter is tied to a brand, a share or a rank. Each vendor can request its own AI Access Report.

  • Brands are spelt as each company writes them on its own website, checked on 7 October 2026: crm2, totalCRM, popcorn CRM and monday CRM begin in lower case and Copper CRM carries its second word. Bigin is a Zoho product and has its own row because the engines name it separately from Zoho CRM.


Running a GEO study for your own category

This study covers CRM software for small businesses in the UK, and the same measurement runs for any category, in any market and language, on the same two halves: what the engines say and cite, and what a probe using their crawler names can reach and read.


We bring more than 15 years in marketing, PR and SEO to this work.



Written by Natali Smith, co-founder of AI, TELL ME! (aipleasetellme.com), 8 October 2026.


bottom of page