How ChatGPT, Claude, and Perplexity Pick Which SaaS to Recommend: Data From 3,334 Citations

Most B2B SaaS founders treat "AI optimization" as a single strategy. They assume that what works for ChatGPT will work for Claude and Perplexity too. Our data shows that assumption is wrong — and it is costing brands visibility on at least one platform they could be winning.

ChatGPT, Claude, and Perplexity each have distinct models for deciding which software brands to recommend. Optimizing for one platform does not automatically improve your standing on the others. In some cases, the signals that matter most for one platform barely register on another.

We analyzed 3,334 individual AI citations across 181 B2B SaaS brands tracked through Besible's citation monitoring system. Using logistic regression, we measured how much each brand-level signal contributes to the probability of being cited on each platform. The result is a clear, data-driven picture of how each AI thinks about authority and trust.

The key finding: domain authority dominates ChatGPT, brand mentions and reviews drive Perplexity, and Claude falls in between — rewarding brands with diverse third-party presence rather than any single dominant signal.

What factors do ChatGPT, Claude, and Perplexity use to recommend SaaS brands?

We trained a binary classification model on 181 brands — each one labelled as high-citation or low-citation based on a median split — and extracted the normalised feature weights for each platform. The result is a clear signal ranking for each AI.

The three platforms have completely different #1 signals:

These aren't minor differences. The signal that matters most for Perplexity barely registers for ChatGPT, and vice versa. One optimisation strategy won't cover all three.

How does ChatGPT decide which SaaS to recommend?

ChatGPT is the most authority-dependent of the three platforms. Domain authority is its #1 signal by a significant margin — roughly twice as influential as its next factor (web mentions). Together, those two signals account for the majority of what separates high-citation from low-citation brands on ChatGPT.

Key finding: ChatGPT's citation model is heavily concentrated at the top. Two signals do most of the work. The other four are secondary.

The mechanism is parametric memory. ChatGPT's responses draw heavily from patterns encoded during training, not real-time retrieval. Training data is biased toward brands that appeared frequently across high-authority publications and well-linked pages. Brands with strong backlink profiles earned more coverage in that training corpus, and that advantage compounds every model update.

The practical implication: ChatGPT is the hardest platform to move quickly. Domain authority takes months to years to build. There's no shortcut. If your competitors already outrank you on DA, you'll see that in ChatGPT citations before anywhere else.

For brands asking "why does ChatGPT recommend my competitors," DA is usually the answer.

How does Claude decide which SaaS to recommend?

Claude uses the most balanced signal set of the three platforms. No single factor dominates.

Domain authority is still Claude's #1 signal, but its advantage over the others is much smaller than on ChatGPT. Brand mentions, web mentions, and reviews all register meaningfully — meaning a brand with broad third-party presence can compensate for a DA gap in ways it simply can't on ChatGPT.

Key finding: The gap between Claude's #1 and #2 factors is far smaller than on ChatGPT. Claude rewards breadth; ChatGPT rewards depth in one specific signal.

Reviews are weighted noticeably higher by Claude than by ChatGPT. A solid G2 or Capterra profile isn't just a nice-to-have for Claude — it's a real input into whether you get cited.

The practical implication: Claude rewards diversification. Backlinks, review platforms, industry mentions, and community discussions all count. A brand that's done most things reasonably well will often outperform a brand that's done one thing exceptionally well.

How does Perplexity decide which SaaS to recommend?

Perplexity is the most distinctive of the three platforms — and the most accessible for smaller B2B SaaS brands.

The signal ranking is almost a mirror image of ChatGPT. On ChatGPT, domain authority is the dominant signal and brand mentions barely register. On Perplexity, brand mentions are the #1 signal and DA drops to third. Reviews — which are a secondary input for ChatGPT — become the second most important factor for Perplexity.

Key finding: Perplexity's top two signals (brand mentions and reviews) are both community-driven, not authority-driven. This is a fundamentally different model from ChatGPT.

This pattern is explained by how Perplexity retrieves information. Unlike ChatGPT, which relies heavily on training-time parametric memory, Perplexity performs real-time web searches to generate responses. It actively crawls forums, review platforms, and community discussions when answering queries. A brand with strong Reddit presence, active review profiles on G2 or Capterra, and recurring mentions in niche communities will surface in those searches even if its DA is modest.

Our research on Reddit and AI citations found that 46.7% of Perplexity citations come from Reddit sources. That concentration is a direct consequence of Perplexity's real-time retrieval model and the weight it places on community-driven brand mentions.

The practical implication: Perplexity is the fastest platform to influence, particularly for newer brands. You do not need years of backlink building to improve your standing here. Getting mentioned in relevant subreddits, building a review base on G2 or Capterra, and earning brand mentions in industry forums can move your Perplexity citation rate in weeks, not months.

What about content quality? Does writing better content get you more AI citations?

Content quality has essentially zero brand-level effect on AI citations. In our model, content quality, content freshness, and structured data quality all land near zero — they don't separate high-citation brands from low-citation ones.

Every brand authority signal (DA, web mentions, reviews, brand mentions) shows a meaningful positive relationship with citation rates. Content-level signals don't.

Key finding: The factors that differentiate which brands get cited are all authority and presence signals. None of them are content quality metrics.

This is counterintuitive, so it is worth explaining clearly. Content quality is not irrelevant — it operates at the page level, not the brand level. Better content helps your pages get retrieved and included in the AI's candidate pool through RAG (Retrieval-Augmented Generation). That matters for individual page citations.

But which brands get cited across hundreds of queries, consistently and repeatedly, is determined by brand authority signals. Think of it as a two-stage process. Stage 1: your content quality determines whether a page is considered. Stage 2: your brand authority signals determine whether you are actually recommended.

If your brand authority is weak, even excellent content will get bypassed in favor of a competitor with stronger DA, more mentions, or more reviews. Content gets you into the pool. Authority gets you chosen.

This does not mean you should write bad content. It means treating content creation and brand amplification as equally important. Publishing an article and then doing nothing to distribute, earn mentions, or drive discussion around it is leaving most of its potential citation value unrealized.

Which AI platform should B2B SaaS brands focus on first?

Start with Perplexity for fastest results. Build toward ChatGPT for long-term reach. Treat Claude as the natural middle ground that responds well to a balanced approach.

Three reasons to prioritize Perplexity first.

Speed. Perplexity's real-time retrieval means new sources can influence citations within days or weeks of appearing. A Reddit thread, a G2 review, or a product mention in a niche community can start contributing to Perplexity citations faster than any other input you can control.

Fairness to smaller brands. ChatGPT's DA dominance means larger, older brands with years of backlink accumulation have a structural advantage. Perplexity's community-signal model is more responsive to current brand activity. A newer brand can realistically compete for Perplexity citations if it has strong community presence, even against competitors with higher DA.

Compounding. Brand mentions earned now for Perplexity become web presence signals that eventually help Claude and ChatGPT. The work is not wasted — it builds the broader signal base that all three platforms read.

ChatGPT reach is worth pursuing, but it requires patient investment in backlinks and coverage from high-DA publications. There is no shortcut. Brands that start building those signals now will see results in 12 to 24 months.

What does this mean for your B2B SaaS AI visibility strategy?

Three things the data makes clear.

The platforms are different enough that one strategy won't cover all three. A brand that only invests in DA and backlinks is optimising for ChatGPT while leaving Perplexity and Claude underserved. A brand that only builds community presence is doing the opposite. The right balance depends on where your brand currently stands — and which gaps are costing you the most citations.

The signals that matter most aren't equally hard to move. Domain authority — ChatGPT's dominant signal — is slow to build and reflects years of accumulated history. Brand mentions and reviews — Perplexity's top two signals — are things a brand can actively improve. This asymmetry matters: it means there are levers you can pull now, and others you're building toward over time.

Knowing the general model isn't the same as knowing your position. This article shows how the three platforms work. What it can't show is where your brand specifically sits on each of these signals — whether your review presence is above or below the threshold that correlates with a +21% citation lift, or how your DA compares to the brands currently being cited in your category.

That's what the audit shows. If you want to understand what GEO is and why it matters for B2B SaaS, start there first — then run the audit to see your brand's actual position across all three platforms.

Want to see how your brand scores on these signals, broken down by ChatGPT, Claude, and Perplexity? Run a free audit at besible.com.

Frequently Asked Questions

What is the single most important factor for ChatGPT citations?

Domain authority is the single most important factor for ChatGPT citations — by a significant margin. In our binary classification model trained on 181 B2B SaaS brands, DA is roughly twice as influential as the next factor (web mentions). ChatGPT relies heavily on parametric memory built during training, which is biased toward brands with strong backlink profiles and broad web presence.

Why does Perplexity cite different brands than ChatGPT?

Perplexity performs real-time web searches to generate responses, while ChatGPT draws more heavily from training-time memory. Perplexity actively crawls forums, review platforms, and community discussions when answering queries. Brand mentions are its #1 citation signal — the reverse of ChatGPT, where brand mentions are near the bottom. A brand with strong community presence but modest domain authority can outperform on Perplexity even while underperforming on ChatGPT.

Does improving my website's content quality help with AI citations?

Not at the brand level. In our model, content quality, content freshness, and structured data quality all show near-zero effects on which brands get cited. Content quality helps at the page level by improving your chances of being retrieved, but brand authority signals — domain authority, web mentions, reviews, and brand mentions — are what determine which brands actually get recommended.

How long does it take to improve AI citation rates?

It depends on the platform. Perplexity responds fastest — new brand mentions, reviews, and community discussions can start contributing within weeks because it uses real-time retrieval. Claude operates on a medium timeline. ChatGPT is slowest because it relies heavily on domain authority, which builds over months to years. Brands focused on fast results should prioritise Perplexity tactics first.

What is domain authority and why does it matter for AI visibility?

Domain authority is a measure of a website's backlink strength — how many other sites link to it and how authoritative those linking sites are. It matters for AI visibility because AI systems, particularly ChatGPT, use DA as a proxy for credibility. Brands with high DA have historically appeared more often across high-quality web sources, which means they're better represented in training data and get cited more. It's the single strongest predictor of AI citations in our dataset across all three platforms.

How does Perplexity use Reddit for SaaS recommendations?

Perplexity's real-time retrieval actively pulls from Reddit threads when answering SaaS queries. In our dataset, 46.7% of Perplexity citations come from Reddit sources — a direct result of Perplexity's community-signal model and its ability to index discussions in real time. Brands with active threads in relevant subreddits, and positive mentions in product comparison threads, see meaningfully higher Perplexity citation rates.