Nearly every answer engine optimisation playbook for SaaS opens the same way: claim your G2 profile, complete it, run a steady review-generation motion. Our own piece on citation-source concentration called review platforms “the highest-leverage surface you directly control” — advice that holds if you have a shipped product and customers to review it.

But a large slice of SaaS doesn’t. If you’re pre-launch, in closed beta, or running a waitlist (Zian is in waitlist beta right now, so this is the playbook we can actually run ourselves), the review-site play isn’t merely hard — parts of it are structurally closed to you. This post maps what’s left: the citation surfaces still open before you have a review profile, the alternatives that accept pre-launch products, and how they compare on effort, time-to-citation and durability.

First, the claim doing the rounds — and where it falls apart

A figure circulating in AEO threads and agency decks says review sites and forums — G2, Capterra, TrustRadius, Reddit — drive roughly half of all AI citations for B2B SaaS. Before building a strategy on it, we tried to trace it to an owning study. It dead-ends, in an instructive way.

The closest candidate origin we found is OtterlyAI’s 2026 AI Citation Economy report (1M+ citations across ChatGPT, Perplexity and Google AI Overviews, January–February 2026). Its summary bullet reads: “Community platforms (Reddit, Quora) capture 52.5% of citations vs. 47.5% for brand domains.” But the same report’s own detailed breakdown says the opposite: “brands represent 52.5% of all citations across AI search engines,” with news sites at 20.3% and community forums at 5.9% (rising to 16.9% on Perplexity). The headline bullet contradicts the body data, the study isn’t B2B-SaaS-specific, and it doesn’t mention G2, Capterra or TrustRadius at all. A garbled summary line, lifted into roundups, is exactly how a “review sites plus forums are half of citations” factoid gets born.

What the owner-published, B2B-SaaS-specific data actually shows is different. Foundation Marketing’s study with AirOps — The Hidden Selection Phase — tracked 5.1 million AI responses and 57.2 million citations over 60 days (December 2025 to February 2026) for 50 B2B brands across seven verticals, on ChatGPT, Gemini, Perplexity, Google AI Mode and AI Overviews. Its findings:

• Only 10.15% of citations linked to brand-owned domains — roughly 90% went to third-party sources.
• Reddit accounted for 20.8% of external citations, the single largest source in six of seven verticals, rising to 30.9% on unbranded discovery queries.
• G2 — in Foundation’s words, “the platform many B2B brands invest in heavily to influence buyer behavior” — accounted for just 4%.

Add those together and review sites plus Reddit land near a quarter of external citations in the one dataset built for this question — significant, but nowhere near half, with the review-site share a rounding error next to Reddit’s. The qualitative core survives: third-party surfaces dominate AI citations and concentrate hard. The “~half” figure belongs to nobody, so we won’t use it — and you shouldn’t either.

The review-site gate is real: G2 excludes beta products by policy

Here’s the part most playbooks skip. It’s not just that a pre-launch product will struggle to gather reviews — G2 won’t list you at all. G2’s own documentation states: “G2 does not accept business-to-consumer (B2C) products or products that are currently in the alpha or beta stage of development” (G2 documentation). If you’re in beta, the “highest-leverage directly editable surface” has a sign on the door.

Capterra’s vendor pages publish no equivalent beta exclusion that we could find, but the deeper constraint is universal: review platforms are corpora of customer testimony, and a waitlist product has no public customers to testify. An empty profile earns no citations. The honest conclusion for early-stage teams is not “try harder on G2” — it’s “spend those months on the surfaces that are actually open.”

Review-site alternatives that are open pre-launch

Two launch-discovery platforms explicitly accommodate products that haven’t shipped:

• Product Hunt. You can schedule a launch for a future date, and Product Hunt’s help centre notes that “if the product is still in beta, or not fully released yet, you can select the option to indicate that it is not available yet” (Product Hunt Help Center). Scheduled launches also get a teaser in the Coming Soon feed, where people can opt into launch-day notification.
• BetaList. A directory built specifically for pre-launch and recently launched startups — its submission guidelines require your own domain and a clear value proposition, not a shipped product.

Be honest about what these are: launch-visibility surfaces, not review corpora. They put your name and category description onto indexed third-party pages — useful for entity establishment — but they won’t carry your citation share the way a mature review profile might later. Treat them as a week of effort, not a strategy.

The six citation surfaces that remain — and how they compare

1. Own-domain depth content. The foundation. Foundation/AirOps found that when a buyer searches a brand by name, the brand’s own content appears in 77.6% of responses — versus brand-owned domains earning just 2.2% of citations on unbranded discovery queries. Translation: your own pages are how engines verify and describe you once anything surfaces your name — so answer-shaped pages, phrased as the questions buyers actually ask, are non-negotiable even though they won’t win discovery queries alone.

2. Comparison pages you publish yourself. Engines assembling “best X” and “X vs Y” answers cite substantive vendor-published comparisons, especially when third-party roundups in a young category are thin. A transparent methodology and honest treatment of competitors matter more here than anywhere — a puff piece reads as marketing and gets skipped.

3. Community answers. Reddit’s 20.8% share is the reason this surface can’t be ignored, and it’s open to anyone with expertise — no shipped product required. The rules are stricter for you, though: pre-launch, you have no users whose experience you could even pretend to represent, so the only viable posture is disclosed, genuinely helpful participation as a practitioner. Our Reddit playbook for B2B SaaS covers the tactics and the failure modes in full; everything in it applies double before launch.

4. Syndication with canonicals. Republishing your best pieces on platforms that support canonical URLs (Dev.to, Hashnode, Medium) puts your content on higher-authority domains engines already crawl, while the canonical keeps ownership pointing home. The trade-off: citations may credit the platform, not you — treat it as distribution for ideas and entity mentions, not a substitute for owned pages.

5. Documentation. A beta product can still publish public docs, and documentation is one of the most retrieval-friendly formats there is: question-shaped headings, concrete specifics, no marketing register. Public docs also do double duty — they’re the material engines use to describe what your product actually does when someone asks about you by name.

6. Data and original research. The strongest pre-launch lever, and the only one that earns editorial citations without customers. Run a real analysis — a benchmark, a dataset, a traced claim (this post’s audit of the “half of citations” figure is the genre) — and you give journalists and community threads something concrete to cite. Notice that nearly every source in this article sells something else entirely; their research is why they get cited.

Comparison: citation surfaces open without a review profile

Time-to-citation has no owner-published benchmark for new domains, so the column below is our ordinal judgement — bands, not measurements.

Surface	Open pre-launch?	Effort	Time-to-citation	Durability
Own-domain depth content	Yes	High, ongoing	Months	High — pages you control, compounding
Self-published comparison pages	Yes	Medium	Months	High if maintained; decays if stale
Community answers (Reddit etc.)	Yes — as a disclosed practitioner	High, ongoing	Months (account credibility first)	Medium — engine- and licensing-dependent
Syndication with canonicals	Yes	Low (reuses existing work)	Weeks–months	Low–medium — platform gets much of the credit
Public documentation	Yes	Medium (you need it anyway)	Months	High — versioned, specific, rarely contested
Original data / research	Yes	High, episodic	Weeks–months after pickup	High while the data stays current
Launch directories (Product Hunt, BetaList)	Yes — explicitly	Low, one-off	Days–weeks (mentions, not answers)	Low — launch-moment visibility
Review platforms (G2 etc.)	No — G2 excludes alpha/beta products	—	—	— (revisit at general availability)

The pattern: everything open to you pre-launch is either slow and durable (owned content, docs, research) or fast and shallow (directories, syndication). There is no fast, durable, pre-launch citation surface. Plan in quarters.

How this fits with what we’ve published before

Two earlier pieces frame this one. Citation-source concentration maps which domains engines trust, assuming you can claim a review profile — the established-vendor view. The Reddit playbook goes deep on the single biggest third-party surface. This post is the missing third leg: what to do before the review-platform door opens, which is where waitlist-stage products — ours included — actually live.

Frequently asked questions

Can a beta or pre-launch SaaS product get listed on G2?

No. G2’s own documentation states it “does not accept business-to-consumer (B2C) products or products that are currently in the alpha or beta stage of development.” A product in beta cannot be listed, and therefore cannot gather G2 reviews or G2 citations, until it reaches general availability. Plan your review-platform motion for launch, not before.

Do review sites and forums really drive half of B2B SaaS AI citations?

No study we could find actually says this. The likeliest origin is a summary line in OtterlyAI’s 2026 report that its own body data contradicts (community forums at 5.9% of all citations, brands at 52.5%). The B2B-SaaS-specific Foundation/AirOps dataset puts Reddit at 20.8% and G2 at about 4% of external citations — heavy third-party concentration, but nothing that supports “half from reviews and forums.”

Which surface should a two-person pre-launch team start with?

Own-domain answer content plus public documentation, because both are assets you need regardless and they’re what engines read when anyone asks about you by name. Add one original-research piece per quarter — it’s the only pre-launch lever that earns citations from domains more trusted than yours. Directories like Product Hunt’s Coming Soon feed are worth the afternoon they take, no more.

Does syndicating posts to Dev.to or Medium hurt the original page?

Not if you set the canonical URL, which those platforms support — the canonical tells crawlers your domain owns the content. The realistic cost is different: when an engine cites the syndicated copy, the citation credit accrues to the platform’s domain. Use syndication to spread ideas and brand mentions while your own domain is young, and keep your best exclusives at home.

How long before AI engines cite a brand-new domain?

There’s no owner-published benchmark, so be sceptical of anyone quoting a precise number. Expect months: engines need to crawl you, and discovery-query citations skew heavily to established third-party surfaces — in Foundation/AirOps data only 2.2% of unbranded-query citations went to brand-owned domains. Measure citation rate over time on your category prompts rather than waiting for a single milestone.

Sources

Figure / claim	Owning organisation	Verified at
5.1M responses, 57.2M citations, 60 days (Dec 2025–Feb 2026), 50 brands, 7 verticals, 5 platforms; 10.15% of citations brand-owned; 2.2% of citations brand-owned on unbranded queries; brand content appears in 77.6% of responses on branded queries; G2 ≈4% of external citations	Foundation Marketing × AirOps	foundationinc.co/lab/foundation-marketing-airops-report
Reddit 20.8% of external citations; 30.9% on unbranded discovery queries	Foundation Marketing × AirOps	foundationinc.co/lab/reddit-ai-citations
Summary line “community platforms capture 52.5% of citations vs 47.5% for brand domains” vs body data “brands represent 52.5% of all citations,” news 20.3%, community forums 5.9% (16.9% on Perplexity); 1M+ citations, Jan–Feb 2026	OtterlyAI	otterly.ai/blog/the-ai-citations-report-2026
“G2 does not accept business-to-consumer (B2C) products or products that are currently in the alpha or beta stage of development”	G2	documentation.g2.com/help/docs/finding-or-listing-a-product-on-g2
Launches can be scheduled in advance; beta / not-yet-released products can mark themselves “not available yet”	Product Hunt	help.producthunt.com/en/articles/479557-how-to-post-a-product

The review-site door opens at launch; the citation clock starts whenever you do. If autonomous AI sales agents are on your evaluation list, we’d rather show you than tell you — Apply For Partnership and see the beta from the inside.