Which SaaS get customers from ChatGPT now?

Is ChatGPT actually sending paying customers to SaaS now?

ChatGPT is already sending real signups, leads and paying customers to SaaS companies today, although only a small number of companies have published enough data to prove it properly.

Ahrefs gives us one of the clearest organic examples. In a 30-day analysis of its own Web Analytics data, AI search generated only 0.5% of visits but 12.1% of signups. Ahrefs said the vast majority of those AI signups came from ChatGPT. Visitors from AI search converted 23 times better than visitors from traditional organic search in that dataset.

Arvow has reported an even higher share. Using PostHog and ChatGPT-tagged inbound links, the AI SEO SaaS says ChatGPT went from 6.7% of tracked signups to 38.1% four months later.

At the other end of the market, NovelCanon provides a much smaller but unusually clean example. The founder of the recently launched writing SaaS checked PostHog after receiving two paid subscriptions and found that the first paying customer had originally arrived from ChatGPT.

Paid acquisition now adds stronger evidence. Jotform says more than 10% of its ChatGPT Ads clicks became signups. Stream says ChatGPT Ads generated signups, sales opportunities and roughly 2x return on ad spend.

We therefore have several different forms of proof: organic signups at scale, attributed paid subscriptions, B2B pipeline and paid campaigns with measurable acquisition economics. The remaining uncertainty is how common these outcomes are across SaaS as a whole.

Why is it so hard to know how many SaaS customers really come from ChatGPT?

ChatGPT almost certainly influences more SaaS purchases than referral dashboards can see.

Imagine someone asks ChatGPT which CRM works best for a five-person sales team using Gmail. ChatGPT recommends three products. The user reads the answer, closes ChatGPT, searches one brand on Google that evening and signs up. Google may receive the final attribution even though ChatGPT introduced the product.

The same problem appears when users copy links, switch devices or type a brand name directly after seeing it mentioned. Some ChatGPT-assisted discovery therefore ends up inside direct, branded-search or unknown traffic.

Semrush's 17-month U.S. clickstream analysis shows why referral data captures only part of ChatGPT usage. ChatGPT outbound referral traffic rose 206% from January 2025 to January 2026, yet ChatGPT used web search on only 34.5% of queries in Semrush's February 2026 sample. A large share of conversations never creates an outbound web-search journey at all.

That forces us to use different confidence levels. A ChatGPT referral click is easy to observe. A signup tied to that referral is much better evidence. A payment or qualified opportunity tied to the original ChatGPT visit is better again.

Companies such as Ahrefs, Arvow, Stream and NovelCanon stand out mainly because they followed the user far enough down the funnel to show something commercially useful.

Which SaaS have the clearest evidence of getting customers from ChatGPT today?

Ahrefs, Arvow, Jotform and Stream currently provide the strongest named SaaS examples we found, while NovelCanon offers a useful small-startup case and LegalNature shows unusually deep intent further down the funnel.

The quality of evidence varies a lot. Ahrefs disclosed traffic and signup shares. Arvow tied signups to tracked ChatGPT links in PostHog. Jotform compared ChatGPT acquisition costs with paid search and paid social. Stream followed ChatGPT users into pipeline and return on ad spend. NovelCanon's evidence is one founder tracing an individual paid subscription, so it tells us much less about scale.

LegalNature sits somewhere in between. Around 40% of visitors from its ChatGPT Ads entered its document-building flow immediately, while roughly 8% completed an entire guided legal document. Those are strong product-engagement events, although LegalNature has not publicly disclosed how many of those users ultimately paid.

“ChatGPT mentioned our brand” is weak evidence. “ChatGPT produced 12.1% of signups” or “this campaign generated sales opportunities and 2x ROAS” answers a much more useful question.

SaaS	What ChatGPT produced	Strength of evidence
Ahrefs	12.1% of signups from AI search; most attributed to ChatGPT	Strong first-party signup data
Arvow	38.1% of tracked signups from ChatGPT	Strong attribution, smaller company
Jotform	10%+ click-to-signup rate; lower signup cost than paid search/social	Strong paid-acquisition data
Stream	Signups, qualified pipeline and ~2x ROAS	Strong B2B commercial data
LegalNature	~40% started document flow; ~8% completed one	Strong intent, payment rate undisclosed
NovelCanon	First paid subscription traced to ChatGPT	Clear customer proof, tiny sample

Is Ahrefs really getting that many signups from ChatGPT?

Yes. Ahrefs found that AI search produced 12.1% of its signups from only 0.5% of visits, with ChatGPT responsible for most of those signups.

The gap between those two numbers is the interesting part. AI's share of Ahrefs signups was roughly 24 times its share of visits. Using the underlying conversion rates, Ahrefs calculated that AI-search visitors converted 23 times better than traditional organic-search visitors.

Their behavior also looked unusually commercial. About 80% of AI visitors landed on the Ahrefs homepage, free tools or product pages, and they viewed 50% more pages than traditional search visitors.

One metric initially looks contradictory: AI visitors spent less time on the site. Yet they still converted far more often. A plausible reading is that many arrived after ChatGPT had already done part of the research. Someone who has spent ten minutes asking which SEO platform fits a particular need may need less time browsing once they reach Ahrefs.

We should keep the 23x figure tied to Ahrefs. There is no good evidence that the average SaaS should expect anything close to that conversion lift.

Still, Ahrefs establishes something important with unusually clean numbers: even tiny ChatGPT referral volume can produce a meaningful share of SaaS signups.

Can Arvow really get 38% of its signups from ChatGPT?

Arvow says 38.1% of its tracked signups came from ChatGPT at its latest reported peak, up from 6.7% four months earlier.

The company tracked inbound URLs carrying a ChatGPT source tag and connected those visits to account creation in PostHog. The reported share rose from 6.7% to 38.1%, a roughly 5.7-fold increase.

That is an exceptional number, so it deserves more skepticism than Ahrefs' result. Arvow sells AI SEO software and publishes case studies about generating customers through AI search. Success in ChatGPT is part of what the company sells.

The underlying measurement is still more useful than a visibility score. Arvow is counting signups associated with ChatGPT-tagged traffic, rather than estimating how often its name appears in AI answers.

Arvow also says most of the growth came from familiar work: SEO, publishing around relevant search intent and building brand authority. Its own case study argues against the idea that one technical trick suddenly unlocked ChatGPT.

We would therefore treat 38.1% as a credible first-party claim rather than a general benchmark. It shows how large ChatGPT can become for a SaaS whose category is frequently researched through AI, especially when the company itself is closely associated with that category.

Can a tiny SaaS get its first paying customer from ChatGPT?

Yes. NovelCanon shows that a new niche SaaS can receive a paying customer from ChatGPT before building much conventional distribution.

NovelCanon is a writing application designed to help novelists keep characters, timelines, world-building details and plot information consistent across long manuscripts.

Its founder said the product had been live for only a few weeks and had barely been promoted when two paid subscriptions arrived on the same day. Looking through PostHog, the founder found that the first customer's original visit came from ChatGPT.

One subscription cannot tell us whether NovelCanon has a repeatable acquisition channel. It does show something that Ahrefs cannot: an established web brand is not required for ChatGPT to produce a customer.

The specificity of the product probably helps. “Software that catches contradictions in a long novel while preserving a story bible and plot threads” maps cleanly to a problem someone might describe in a ChatGPT conversation.

Large brands have an authority advantage. Very focused startups have another route: being unusually relevant to a narrow question.

Does ChatGPT traffic usually convert better than Google traffic?

ChatGPT visitors often look more valuable than ordinary search visitors, but the current evidence is too uneven to give SaaS a universal conversion multiplier.

Ahrefs measured a 23x conversion-rate advantage over organic search. That is the spectacular case, not the baseline.

What we see elsewhere is directionally similar but far less extreme. Recent industry studies have repeatedly found that AI-referred visitors arrive with stronger commercial behavior than ordinary web traffic. Similarweb's latest cross-industry analysis says AI referral visits more than doubled year over year, while Adobe found AI-sourced U.S. retail traffic converting 54% better than non-AI traffic by May 2026.

Retail is obviously different from SaaS, so we cannot transfer that 54% directly. The pattern still fits what SaaS companies such as Ahrefs, Jotform and Stream are observing: people arriving from a conversation often know more about the product and their own problem before clicking.

Jotform offers a useful paid comparison. More than 10% of its ChatGPT Ads clicks converted into signups, and its cost per signup was around 30% lower than paid search and 60% lower than paid social during the reported campaign period.

Stream saw the same quality difference earlier in the funnel. More than 70% of ChatGPT Ads clicks became actual website sessions, versus below 10% across the other channels in its controlled test.

So the direction looks real before the benchmark does. ChatGPT can produce unusually high-intent SaaS traffic. How much better it converts depends heavily on the product, query and way attribution is measured.

If ChatGPT visitors convert so well, why isn't ChatGPT already huge for SaaS?

ChatGPT still sends far fewer website visits than Google, so even excellent conversion rates often sit on top of a small traffic base.

Ahrefs captures the issue perfectly. AI search generated 12.1% of signups, which sounds enormous, while generating only 0.5% of visits.

At web scale, AI referral traffic is growing very quickly but still represents a low-single-digit share for most websites. Similarweb's latest industry study measured an average 770.7 million monthly AI referral visits worldwide between June 2025 and May 2026, up 117.4% year over year. Every industry it tracked at least doubled.

Semrush independently found ChatGPT outbound referrals increasing 206% over a one-year period.

The growth rates are huge partly because the starting point was small. That is why statements such as “AI search is replacing Google” run well ahead of the data.

The more useful comparison is customers produced per visit. A SaaS could receive only 1% of its traffic from ChatGPT and still care a lot about the channel if those visitors generate 5%, 10% or more of new accounts.

Ahrefs already fits that pattern.

Has ChatGPT recently become better at sending people to SaaS websites?

Yes. ChatGPT made a visible jump as an outbound traffic source this year, and two large independent datasets picked up the change.

Similarweb studied desktop referral behavior around a sudden May 2026 shift in ChatGPT outbound traffic. Total ChatGPT referral traffic jumped 157.7% week over week. Visits going directly to brand homepages rose 354.7%.

The destination changed as well. Before the shift, roughly 26–32% of ChatGPT referrals landed on homepages. Afterward, the share jumped to around 60% and stayed close to that level during Similarweb's observed period.

SE Ranking saw a similar jump in a much larger international sample. Across 101,574 websites in 250 countries and territories, ChatGPT's worldwide referral share rose from 0.23% to 0.32% in one month, a 36.7% increase. The previous four months had produced only about 1.5% growth combined.

Those datasets do not prove the exact product feature that caused the change. The numerical shift is much firmer than any single explanation for why it happened.

For SaaS, homepage traffic is particularly interesting. A user who leaves ChatGPT for a company's root domain is often researching the company itself rather than reading one informational article.

Measure	Before the shift	After the shift
Total ChatGPT referrals	Previous baseline	+157.7% week over week
Homepage referrals	Previous baseline	+354.7% week over week
Share landing on homepages	~26–32%	~60%
Worldwide ChatGPT referral share in SE Ranking sample	0.23%	0.32%

Are ChatGPT Ads already getting customers for SaaS?

Yes. Jotform and Stream now provide direct evidence that paid ChatGPT acquisition can generate SaaS signups and commercially useful pipeline.

Jotform's OpenAI case study is especially useful because it compares channels. More than 10% of ChatGPT Ads clicks converted into signups. Cost per signup came in about 30% below paid search and 60% below paid social. Cost per click was also more than 40% below Jotform's non-search paid media.

The company did not get those results immediately. Jotform says it completed more than 100 campaign optimizations. Use-case-specific landing pages eventually increased signup rate by 2.5 times, while moving from reach campaigns to click campaigns reduced cost per click by four to five times.

Stream tested ChatGPT Ads against other channels with roughly $10,000 of media spend per test group. More than 70% of ChatGPT Ads clicks became website sessions, while the comparison channels remained below 10%. Stream says ChatGPT produced signups, sales opportunities and roughly 2x ROAS.

LegalNature provides another current example. Nearly 40% of visitors from ChatGPT Ads immediately began building a legal document, around 8% completed the full guided document, and 95% of completed documents were finished within one hour of arrival.

The public evidence is still mostly first-party advertiser data published by OpenAI, so we should not treat these numbers as independent market benchmarks. The actual actions being measured are substantial enough to take seriously.

Can expensive B2B SaaS get leads from ChatGPT too?

Yes. ChatGPT is already producing qualified leads for B2B software priced in the thousands or tens of thousands of dollars.

Stream is the strongest named example because its ChatGPT campaigns produced actual sales opportunities rather than stopping at registrations.

InterTeam Marketing adds breadth. The agency says it has generated more than 150 qualified leads through ChatGPT Ads across three B2B SaaS clients, with cost per lead around 60% below Google. The campaigns covered AI software, tax SaaS, procurement software and project-management products.

The products ranged from roughly $3,000–$10,000 offers to software costing more than $50,000. InterTeam has not named the three companies, which limits how independently we can verify the claim, but the pricing range is still informative.

For expensive SaaS, the most interesting opportunity may sit above the final purchase. Buyers ask about integrations, migration, security, pricing, competitors and whether a product can handle a particular workflow. Those are natural ChatGPT questions.

ChatGPT can therefore help create and qualify the shortlist before a salesperson ever enters the process.

Which kinds of SaaS are best placed to get customers from ChatGPT now?

SaaS products with a clear problem, an explainable use case and several realistic alternatives appear best placed to win customers from ChatGPT today.

Developer software fits extremely well. Someone can ask how to add video calls to an app, which authentication provider works with a particular stack or which database suits a workload. Stream's results show that those conversations can move directly into pipeline.

Marketing software has the same property. Ahrefs and Arvow benefit because users frequently compare SEO, content and analytics tools before choosing one.

Jotform shows how broad SaaS can still work when individual use cases are concrete. “I need software to create an employee application form with e-signature and approval steps” gives ChatGPT much more to work with than “What software should my company buy?”

Vertical SaaS may have an even cleaner fit. NovelCanon focuses on consistency in long novels. LegalNature focuses on specific legal documents. A narrow problem makes it easier for ChatGPT to explain why one product belongs in the answer.

We would therefore focus less on whether a company calls itself “AI SaaS” and more on whether buyers naturally describe their requirements before purchasing.

The easier the product is to recommend from a conversation, the stronger the fit.

Are ChatGPT apps creating a second SaaS acquisition channel?

Yes. SaaS products can now appear inside ChatGPT itself, creating a distribution path that normal referral analytics barely captures.

ChatGPT supports external apps and plugins that can perform work without requiring the user to begin on the SaaS company's own website. OpenAI lists products including Canva and Figma among its ChatGPT apps, while its business plugin catalog includes software such as Lovable.

Lovable is a good example of how different this funnel can become. A user can ask Lovable inside ChatGPT to build an application, modify the application through follow-up prompts and continue working without first discovering Lovable through Google.

OpenAI says app developers can connect their existing backend so customers can log in and access premium functionality.

For acquisition, that removes several traditional steps. Someone may discover the product by invoking it inside a conversation, experience part of the product there and only later create or connect an account.

We do not yet have solid public numbers showing how many new Canva, Figma or Lovable subscribers originate this way. Calling it a major customer-acquisition channel would therefore be premature.

The distribution mechanism itself is already real, and it gives SaaS companies another reason to look beyond chatgpt.com referral traffic when measuring ChatGPT's influence.

Does ranking well on Google help a SaaS get customers from ChatGPT?

Yes. Strong SEO and brand authority still appear to give SaaS companies a major advantage in ChatGPT discovery.

Ahrefs is almost the perfect illustration. Years of guides, product pages, free tools, comparison content and technical authority give ChatGPT abundant material for understanding where Ahrefs fits.

Arvow reached a similar conclusion from its own data. The company says its rise from 6.7% to 38.1% of tracked signups came mainly from classic SEO and brand-building work rather than a special ChatGPT trick.

That does not mean Google rank and ChatGPT visibility move in lockstep. ChatGPT can draw on different sources, summarize several pages and occasionally surface niche products that would struggle to rank first for broad Google queries. NovelCanon shows that a small product can still get through.

But the companies with the deepest public evidence today tend to make themselves extremely easy to understand online. Their pricing, use cases, integrations, templates, documentation and alternatives are explicit.

That work helps humans, search engines and AI models for largely the same reason: there is less ambiguity about what the product actually does.

Which SaaS pages seem to turn ChatGPT discovery into signups?

Product pages, homepages and tightly matched use-case pages currently look more valuable for ChatGPT acquisition than generic informational traffic alone.

Ahrefs found that roughly 80% of its AI visitors landed on its homepage, free tools or product pages.

Similarweb then observed a broader change across the web: after the May referral shift, around 60% of referrals in its sample began landing directly on brand homepages, compared with roughly 26–32% beforehand.

Jotform adds conversion evidence. Its use-case-specific landing-page work increased signup rate by 2.5 times during its ChatGPT Ads experimentation.

These findings fit together well. ChatGPT can answer much of the informational question itself. By the time someone clicks, the job of the website may already have changed from “teach me about this topic” to “show me whether this product actually fits.”

That makes clear product positioning, pricing, integrations, alternatives, templates and specific workflow pages especially useful.

Publishing another generic “What is CRM?” article may help source visibility somewhere upstream. A page answering exactly what a buyer needs before signing up is closer to where the commercial value appears.

Is the hype around GEO and ChatGPT customer acquisition getting ahead of reality?

Yes, some of the GEO hype is far ahead of the evidence, even though ChatGPT customer acquisition itself is already real.

The incentive problem is obvious. Arvow sells AI SEO. Ahrefs sells SEO software. Agencies publishing spectacular ChatGPT growth stories often sell GEO or ChatGPT advertising services. OpenAI's strongest advertising examples appear on OpenAI's own advertising site.

That does not make the underlying figures false. It does mean we should distinguish measurement from marketing.

The strongest claims survive that filter because they contain hard downstream events. Ahrefs reports signups. Jotform reports signup conversion and comparative acquisition costs. Stream reports pipeline and ROAS. NovelCanon identifies a paid subscription. LegalNature measures people completing a long document-building flow.

Broader independent datasets also confirm that something meaningful is changing. Semrush measured ChatGPT referrals up 206% year over year. Similarweb measured worldwide AI referrals up 117.4%. SE Ranking detected a sudden 36.7% monthly jump across more than 100,000 sites.

So two things are happening at once: a genuine acquisition channel is emerging, and an industry has formed around exaggerating how mature that channel already is.

The numbers support the first claim much better than the second.

Could ChatGPT become as important as Google for SaaS acquisition?

ChatGPT is nowhere near Google in total SaaS traffic today, although its unusually high-intent visitors could make it important long before the traffic volumes become comparable.

The Ahrefs case shows why raw traffic share can mislead us. AI search represented only 0.5% of visits but produced 12.1% of signups during its measured period.

Similarweb's latest data says AI platforms averaged 770.7 million monthly referral visits worldwide across the year it studied, up 117.4%. That is substantial web traffic, but AI referrals still make up a low-single-digit share for most sites.

ChatGPT dominates that AI referral pool. Similarweb has measured it at more than 80% of AI referrals among the world's top 1,000 domains in its earlier large-domain analysis.

Google therefore remains vastly larger as a web acquisition engine. We see no evidence today that the average SaaS company should expect ChatGPT to replace it soon.

ChatGPT does not need Google's traffic volume to matter. If a small stream of ChatGPT visitors keeps converting far above baseline, SaaS companies will care about the channel much earlier.

That is already happening at Ahrefs, and paid campaigns from Jotform and Stream suggest the same dynamic can appear outside organic referrals.

Should SaaS companies actively optimize for ChatGPT now?

Yes, SaaS companies whose products are researched and compared before purchase should actively track ChatGPT now, while keeping the investment proportional to the customers it actually produces.

The first job is measurement. A company needs to preserve first-touch source through signup and CRM or payment events rather than stopping at website sessions. Otherwise, a surge in ChatGPT traffic tells us almost nothing about revenue.

The website also needs to explain the product with very little ambiguity. Pricing, integrations, alternatives, technical documentation and concrete use cases help ChatGPT understand the product and help already-informed visitors make a decision quickly.

For paid acquisition, the recent Jotform and Stream results make ChatGPT Ads worth testing for SaaS with clear conversational intent. The platform is young enough that we should expect performance to move around quickly as inventory, targeting and advertiser competition change.

The same restraint applies to GEO spending. A company receiving three ChatGPT visits a month does not need a large “AI search strategy” simply because the category is fashionable. A company discovering that 8% of new trials already begin with ChatGPT has a very different problem.

These days, the better question is less “Should every SaaS do GEO?” and more “How many customers are we already getting from ChatGPT, and what do those customers have in common?”

Which SaaS get customers from ChatGPT now?

Yes, SaaS companies are getting real customers from ChatGPT now, with Ahrefs, Arvow, Jotform and Stream providing the clearest named evidence we found.

Ahrefs gives us the strongest public organic example at meaningful scale: AI search produced 12.1% of signups from 0.5% of visits, and most of those AI signups came from ChatGPT. Arvow reports an even higher 38.1% ChatGPT share among tracked signups, although its smaller size and position inside the GEO industry make that figure harder to generalize.

Jotform and Stream give us the strongest paid evidence. Jotform converted more than 10% of ChatGPT Ads clicks into signups while paying less per signup than on paid search or paid social. Stream followed ChatGPT users further downstream and reported actual sales opportunities plus roughly 2x ROAS.

NovelCanon shows that ChatGPT can also surface a tiny SaaS and produce a paying customer before the company has built much conventional distribution. LegalNature adds evidence that ChatGPT users can arrive remarkably deep in the buying journey, with around 8% of its ad visitors completing a full guided legal document.

The scale remains uneven. Most SaaS companies are still getting far more customers from Google, direct traffic, paid search, affiliates, social channels and existing brand demand. Public proof of ChatGPT-generated revenue is still much rarer than screenshots showing AI mentions.

But the question has changed. We no longer need to ask whether ChatGPT can send SaaS customers at all. It can, and several companies can now show the signups, pipeline and payments.

What we still need to learn is how many SaaS categories can turn that behavior into a large, repeatable acquisition channel. For now, the companies getting the clearest results tend to share one trait: users can describe a specific problem in ChatGPT, and the software provides an obvious answer to it.