Guides / SaaS citations
How to get your SaaS product cited by ChatGPT and Perplexity when buyers research your category
By Olly, founder of AEOscore, AEO consultant and full-stack developer, 20+ years in search11 min read
How do I get my SaaS product cited by ChatGPT and Perplexity?
On this page
The difference
Why is SaaS different from other categories?
Software buyers ask the engines for a shortlist, so a SaaS product has to appear in the lists and comparisons the engines read.
G2's March 2026 survey of 1,076 B2B software buyers found 51% now start their research with an AI chatbot more often than with Google [1], and Ofcom found 54% of UK adults use AI tools such as ChatGPT, Copilot or Gemini (April 2026) [2]. The questions they ask are shaped for a list: "best CRM for a 20-person agency", "Xero vs QuickBooks", "alternatives to HubSpot". In our 304-query study of UK B2B searches, 88% of comparison queries showed a Google AI Overview (17 August 2026). Being cited beats being left out: Seer Interactive found organic click-through of 2.07% on informational queries when the brand was cited in the AI Overview, against about 1.6% when it was not, once one outlier account is excluded (April 2026) [8]. Both sit below the 3.35% seen when no AI Overview appears.
Step 1 of 7
Which questions should you track?
Write down 20 to 40 questions a buyer in your category would type into ChatGPT, in their words, and ask each one on every engine before you change anything.
Mix four kinds: shortlist questions ("best payroll software for a 50-person UK firm"), comparisons ("X vs Y"), problem questions ("how do I stop month-end payroll errors") and questions about you by name. Sales calls, demo requests and support tickets give you the wording; keyword tools give you shorter search phrases. Freeze the list once it is written. On our own site the list of brands named changed between days in 81% of question and engine combinations (18 to 20 August 2026), so the only fair test is the same questions asked the same way every week.
Step 2 of 7
How do you make your own pages easy to quote?
Open each important page with a one or two sentence answer to the question it targets, use the buyer's question as the heading, and put the figures a buyer would quote in visible text.
Perplexity splits documents into short, self-contained passages and ranks each one against the question on its own [4], so a section that only makes sense after the paragraph above it rarely gets used. The GEO study from Princeton and IIT Delhi found that adding citations, quotations and statistics raised a source's share of generative answers by 30 to 40% [3], and Google rankings overlap with it: Ahrefs found 28.6% of the URLs Perplexity cited also ranked in Google's top 10 for the same query [5]. For a SaaS site that means real prices on the pricing page, one page per integration and per use case, plan limits in numbers, and a visible updated date on anything with figures.
Step 3 of 7
How do you get onto the pages the engines quote?
Find the roundups, listicles, review profiles and threads the engines cite for your questions, then get onto them one by one, starting with the pages that already name your rivals.
This is where most citations come from. In 22,065 sources Perplexity cited for three UK businesses we track, 83.4% were third-party pages and 3.5% were the business's own site; "best", "top" and "vs" lists alone made up 20.6% (3 August to 2 November 2026). For the question this page answers, the engines cited 98 sources on 58 different websites over four weekly scans, and YouTube was cited most, 7 times. Email list authors with your facts and a reason to include you, answer threads where you have something useful to add, and keep your G2, Capterra and TrustRadius profiles complete and current.
“Most SaaS sites I look at have decent pages. What they're missing is anyone else saying their name on the lists and threads the engines read.”
Step 4 of 7
Should a SaaS company publish comparison and alternatives pages?
Yes: publish a fair comparison against each rival buyers name alongside you, and an alternatives page for your category leader, because shortlist and comparison questions are where the engines build a list.
Comparison and alternatives pages were 13.4% of the 1,981 times the engines cited our own site, more than our homepage. Write them fairly: say what the rival does better, give prices with the date you checked them, and lead with the reason a buyer would pick you.
“If you only write one page this quarter, make it the comparison against your two biggest rivals, written fairly. Buyers ask for that right before they book a demo.”
Step 5 of 7
How do you make your brand one consistent entity?
Use the same product name, one-line description, company name, address and prices on your site, your review profiles, LinkedIn, Crunchbase and directories, because an engine that finds two versions of a fact may pick either one.
Write the canonical wording once, publish it on a plain facts page, and copy it everywhere. Then hunt down the old versions: directory listings with a retired price, a press page with the product's old name, a partner page with last year's plan names. Plain-fact pages earn their keep here: on our own site, About, terms and methodology pages were the most-cited group of all.
Step 6 of 7
Can the AI crawlers read your site?
Check that robots.txt and your firewall allow the search crawlers, OAI-SearchBot for ChatGPT search and PerplexityBot for Perplexity, and that your key copy is in the HTML before any JavaScript runs.
Each engine uses its own crawler names: OAI-SearchBot fetches pages for ChatGPT search and GPTBot collects training data, PerplexityBot feeds Perplexity's index, Claude-SearchBot fetches pages for Claude's search while ClaudeBot collects training data, and Google-Extended controls whether your content trains and grounds Gemini, which Google says does not affect Search ranking. A bot-protection setting at Cloudflare can block them without anyone changing a line of your site. SaaS marketing sites built as single-page apps carry a second risk: in one 2026 test, Perplexity read the page's raw HTML and never saw what the script added [7].
Step 7 of 7
How do you measure whether it is working?
Re-ask the same frozen questions every week and record three numbers: mention rate (answers that name you), citation rate (answers that link to your site) and share of voice against named rivals.
One check tells you nothing, because answers vary from run to run. Log every answer, take the weekly rates, and report the trend once a month. Judge a fix at week eight and give it until week twelve before you decide it did not work.
Worked result
Does this work for a SaaS product? Our own before and after
On our own product, AEOscore, the share of answers naming us across the same five engines rose from 12.6% on our first full scan (18 August 2026) to 19.3% across four weekly scans from 14 October to 2 November 2026.
| Measure | Before: 18 August 2026 | After: 14 October to 2 November 2026 |
|---|---|---|
| Questions tracked | 40 | 42 |
| Answers scored | 199 | 834 (4 weekly scans) |
| Answers naming AEOscore | 12.6% | 19.3% |
| Answers linking to aeoscore.co.uk | 8.5% | 17.5% |
| Perplexity answers naming AEOscore | 2 of 40 (5%) | 65 of 168 (39%) |
The rise came on Perplexity, which searches the web for almost every question: it named us in 2 of 40 answers in August and 65 of 168 from mid-October, while the other four engines together held level. We shipped 8 targeted fixes between 17 and 24 August, three of them the day before the August scan, then new comparison pages, guides and tips through September and October, the same steps as above. The first week’s fixes, with how fast each engine picked them up, are logged in how long AEO takes.
AEOscore is a UK B2B SaaS, so the case is ours and named rather than an anonymised client. Engines: ChatGPT, Gemini, Perplexity, Google AI Overviews and AI Mode. Claude is left out of both sides so the two windows compare like for like: it was not scanned in August. The panel grew from 40 to 42 questions; rates are per answer.
Timing
How long does it take to show up in AI answers?
Plan on 4 to 12 weeks from a fix going live to mentions you can rely on: judge it at week eight, and give it until week twelve before you decide it did not work.
| When | Do this | What to expect |
|---|---|---|
| Weeks 1 and 2 | Freeze the questions, run the baseline, fix crawler access, rewrite the three pages closest to your best questions. | Crawlers re-read the changed pages. On narrow questions, ChatGPT search and Perplexity can name you within days. |
| Weeks 3 to 6 | Publish the comparison and alternatives pages, complete your review profiles, email the first list authors. | Live-search engines start retrieving the new pages. Google AI Overviews can start to appear from around week four. |
| Weeks 6 to 12 | Keep the list outreach going, answer threads, judge each fix at week eight against the frozen questions. | Repeat mentions across several engines where your pages and third-party sources agree. Broad category questions and answers from training data move last. |
If nothing has moved by week twelve, the usual causes are a crawler block, a page that buries its answer below an introduction, or no third-party mentions at all. The engine-by-engine detail is in how long AEO takes, and why Perplexity cites your competitor covers the Perplexity side in depth.
Tools
Which tools track AI citations, and which make the fixes?
Most AI visibility tools track and report, and AEOscore is the one in this table that also makes the changes, on its £1,495 a month tier.
| Tool | Engines covered | How often | Makes the fixes? |
|---|---|---|---|
| AEOscore (ours) | Six on every tier: ChatGPT, Claude, Gemini, Perplexity, Google AI Overviews and AI Mode | Weekly on Monitor, daily on Monitor + Strategist; 25 questions included, up to 40 | Written fix plans on every tier; the changes made for you at £1,495 a month |
| Peec AI | Three chosen from ChatGPT, AI Mode, AI Overviews, Copilot, Gemini and Naver AI; Perplexity and Claude on Enterprise | Daily | No: recommends |
| Profound | Free trial on ChatGPT, Gemini and AI Overviews; up to nine on Enterprise | Daily on the 7-day trial; Enterprise sized to your prompt set | No: reports, and AI agents draft content |
| Otterly.AI | ChatGPT, AI Overviews, Perplexity and Copilot; Claude, Gemini and AI Mode as paid add-ons | Daily | No: recommends, plus a crawlability audit |
| Semrush AI Visibility Toolkit | ChatGPT, Google AI, Gemini and Perplexity | Daily | No: flags blockers and recommends |
Checked on each vendor’s own pricing page, 2 November 2026. Full comparisons: all tools compared, Profound alternatives and monitoring vs fixing.
FAQ
Getting a SaaS product cited, asked directly
- What is the difference between being mentioned and being cited?
- A mention is your brand named in the answer text; a citation is a link to one of your pages among the answer's sources. They move separately. On our own site, between August and early November, the share of answers naming us rose from 12.6% to 19.3%, and the share linking to us from 8.5% to 17.5%.
- Which pages on a SaaS site get cited most often?
- Pages that state plain facts. Of the 1,981 times the engines cited our own site (17 August to 2 November 2026), About, terms and methodology pages took 14.5%, our UK agency service page 13.7%, comparison pages 13.4%, the homepage 12% and pricing 9.1%. Every page that answers one question clearly gets its share.
- Do review sites like G2 and Capterra matter for AI citations?
- They matter less than lists. Review sites and directories were 2.2% of the sources Perplexity cited in our scans, against 20.6% for "best", "top" and "vs" lists. Keep G2, Capterra and TrustRadius complete anyway: roundup authors and buyers check them, and a stale profile repeats old prices.
- Does ChatGPT work differently from Perplexity here?
- Yes. Perplexity searches the web for almost every question and lists its sources, so a new page can show up within days. ChatGPT searches for some questions and answers the rest from what the model already knows: a searched question can pick up a new page quickly, while the rest lean on third-party coverage built up over months.
- Do I need schema markup to get cited?
- No. Schema is housekeeping: it helps machines read your page details accurately and keeps dates correct. An Ahrefs test of 1,885 pages found no platform showed a meaningful citation increase after schema was added (reported May 2026). Put every fact you want quoted in visible text.
- Can I pay ChatGPT or Perplexity to recommend my SaaS?
- Not a place in the answer itself. You can buy labelled ads shown beside some AI answers: OpenAI began showing ads in ChatGPT in 2026, below the answer and labelled, and says ads do not influence what ChatGPT answers. What the answer says comes from the page, third-party and crawler work above.
- How do I track whether this is working?
- Ask the same frozen questions on a fixed schedule, weekly is enough, and log mention rate, citation rate and share of voice against named rivals. Look at the trend over four or more scans, because a single check can swing either way.
How AEOscore helps
What does AEOscore cost, and what does it do?
AEOscore runs your buyer questions on six engines every week and writes a fix plan for every gap at £99 a month, makes the changes for you at £1,495 a month, and starts with a £395 one-off AI Visibility Audit credited against your first month.
On Monitor your team makes the changes from the plan and we show whether the answers moved; on Monitor + Strategist a named UK expert makes them for you, and every shipped fix is re-checked at 7 and 28 days. See pricing and what monitoring tools leave to you, or start with the free AEO score.
Legal entity: GOREBLIMEY LTD (Goreblimey Ltd), registered in England, company number 07582417, based in Cheltenham, trading since 2011. Goreblimey Ltd is a design and development agency; AEOscore is its product.
This page carries Article and FAQPage structured data and a correct last-updated date, so machines can read its details and its age. That is housekeeping: it does not make the page more likely to be cited.
Sources
Sources
- G2, G2 research: The Answer Economy. Accessed 2 November 2026. Published 15 April 2026, survey of 1,076 B2B software buyers and decision-makers. "Half of B2B software buyers (51%) now start their research with an AI chatbot more often than with Google."
- Ofcom, UK adults' media and online lives revealed. Accessed 2 November 2026. Published 2 April 2026. 54% of UK adults use AI tools such as ChatGPT, Copilot or Gemini.
- Aggarwal et al., Princeton University and IIT Delhi, GEO: Generative Engine Optimization (KDD 2024). Accessed 2 November 2026. First published 16 November 2023, revised 28 June 2024. The Perplexity test used 200 queries with source pages supplied as uploads, not live search, and measures how much of the answer a source earns rather than whether it is cited at all.
- Perplexity Research, Architecting and evaluating an AI-first search API. Accessed 2 November 2026. Published 25 September 2025. Documents are split into self-contained spans ranked at query time; hybrid keyword and semantic retrieval with a reranker; first-cut filters remove stale content; PerplexityBot complies with robots.txt.
- Ahrefs, Only 12% of AI cited URLs rank in Google's top 10 for the original prompt. Accessed 2 November 2026. Published 11 August 2025, 15,000 long-tail prompts. 28.6% of the URLs Perplexity cited ranked in Google's top 10 for the same query.
- Search Engine Journal, Schema markup didn't move AI citations in Ahrefs test. Accessed 2 November 2026. Published 11 May 2026, reporting an Ahrefs test of 1,885 pages: "No platform showed a meaningful citation increase after schema was added." Perplexity was not among the platforms tested.
- Search Engine World, Do AI assistants actually render your JavaScript when grounding? We put it to the test. Accessed 2 November 2026. Published 9 June 2026, updated 2 July 2026. Perplexity reported the page's raw HTML and did not run its script. A single test.
- Seer Interactive, AIO impact on Google CTR: 2026 update. Accessed 2 November 2026. Published 24 April 2026, 53 brands, 5.47M queries. Organic click-through on informational queries was 2.07% when the brand was cited in the AI Overview, 0.94% when it was not (about 1.61% excluding one account Seer flags as skewing that figure), and 3.35% when no AI Overview appeared.
- OpenAI, Our approach to advertising and expanding access to ChatGPT. Accessed 2 November 2026. Published 16 January 2026. Ads shown below the answer, labelled and separate from it; one of OpenAI's five principles is that ads do not influence ChatGPT's answers. The page blocked our fetch, so the wording is checked against press coverage of the post.
AEOscore figures come from our own weekly scans of aeoscore.co.uk and two other UK businesses, read on 2 November 2026. £395 one-off AI Visibility Audit, credited against your first month, £99 a month Monitor, £1,495 a month Monitor + Strategist. Prices exclude VAT.
See which buyer questions name your rivals
The free AEO score asks your buyer questions on all six engines and shows who gets named instead of you, human-checked within 1 to 2 working days. Domain and email, no card.
- ChatGPT
- Claude
- Gemini
- Perplexity
- Google AI Overviews
- Google AI Mode