Industry note
AI Visibility for SaaS Companies: Which Sources AI Engines Actually Cite in Enterprise Software Answers
Index data for enterprise software: across the six SaaS buyer questions, editorial publications hold 4 of the 60 top-ten slots, TechCrunch is rank 190 of 1,375 on the news question and holds no buyer row at all, and TechTarget is the inverse. What AI visibility means for a SaaS company once the sources are measured, and the plan rebuilt on it.
Updated September 23, 2026
AI visibility for a SaaS company is the frequency with which the company appears as a cited source in AI-generated answers to its category's buyer questions. The usual answer to the "how" is a thesis: that TechCrunch and VentureBeat carry the strongest weight for SaaS category queries, with Forbes and Business Insider behind them. The Machine Relations Index has now measured the enterprise software category across six answer engines, and the measurement does not support that thesis.
In the September 19, 2026 release, enterprise software publishes seven question shapes. Across the six buyer shapes, editorial publications hold 4 of the 60 top-ten slots (6.67%). TechCrunch appears on one shape only — the news question — at rank 190 of 1,375 sources, cited in 6 of 596 answer runs (1.01%). TechTarget, which almost no SaaS PR plan targets, is rank 1,234 of 1,375 on that same news question and rank 3 on two of the six buyer questions. The outlets are not weak; they are aimed at the wrong question.
This page covers what actually holds SaaS buyer answers, where the tech press does and does not reach, what a company's own domain can win, and the plan rebuilt on the evidence.
Key Takeaways for SaaS Companies
- Specialist sources, not general tech media, hold the SaaS buying answers. Of the 60 top-ten slots across the six enterprise software buyer shapes, 31 (51.67%) carry the release's Other observed source label — implementers, consultants and category-specific research sites — 11 (18.33%) are vendor-owned domains, and 4 (6.67%) are editorial publications.
- Your own domain is a real lane, and it is shape-dependent. Vendor-owned sources take 34 of the top 100 on news-driven citations and 18 of the top 100 on problem-first research, but only 6 on how buyers choose.
- One database holds three of the six buyer questions. Erpresearch.com is rank 1 on best tools (34 of 101 runs, 33.66%), how buyers choose (28 of 99, 28.28%) and is it worth it (26 of 107, 24.30%).
- The trade outlet beats the tech outlet on buying questions and loses on news. TechTarget is rank 3 on how buyers choose (18 of 99, 18.18%) and rank 3 on comparisons (14 of 114, 12.28%). SaaStr is rank 4 of 1,375 on the news question (48 of 596, 8.05%) and holds no buyer row.
- Measure by question shape or you are measuring nothing. No single source leads all six buyer shapes, so a single "AI visibility score" averages six different competitive fields into one number you cannot act on.
What Holds Enterprise Software AI Answers
Each shape in the Index keeps its own run denominator and its own leaderboard. Here is the rank-1 source for every published enterprise software shape in the September 19, 2026 release.
| Question shape | Rank-1 source | Source role | Cited runs | Rate | Total ranked sources |
|---|---|---|---|---|---|
| Best tools | Erpresearch.com | Market and company database | 34 of 101 | 33.66% | 167 |
| How buyers choose | Erpresearch.com | Market and company database | 28 of 99 | 28.28% | 172 |
| Is it worth it | Erpresearch.com | Market and company database | 26 of 107 | 24.30% | 194 |
| Comparisons | Erpfocus.com | Other observed source | 23 of 114 | 20.18% | 186 |
| Top lists | Procuredesk.com | Other observed source | 19 of 114 | 16.67% | 264 |
| Problem-first research | Community and social platform | 14 of 106 | 13.21% | 258 | |
| News-driven citations | Community and social platform | 66 of 596 | 11.07% | 1,375 |
Read the column of rank-1 sources and the SaaS editorial plan writes itself differently. An ERP research database, an ERP comparison site, a procurement blog, a subreddit and LinkedIn are holding the top slot on the questions a buyer asks before a shortlist exists. None of them is a publication a SaaS communications budget is normally pointed at.
Where the Tech Press Actually Reaches
The four outlets that thesis names are all measured in the same release, and their enterprise software rows are thin and one-sided.
| Outlet | Enterprise software rows | Best enterprise software position | Buyer-question rows |
|---|---|---|---|
| TechCrunch | News-driven only | Rank 190 of 1,375, cited in 6 of 596 runs (1.01%) | None |
| VentureBeat | News-driven only | Rank 400 of 1,375, cited in 3 of 596 runs (0.50%) | None |
| Business Insider | News-driven only | Rank 717 of 1,375, cited in 1 of 596 runs (0.17%) | None |
| WIRED | None in this category | No enterprise software row in this release | None |
| Forbes | Five of the seven shapes | Rank 3 of 167 on best tools, cited in 14 of 101 runs (13.86%) | Four |
TechCrunch is not a weak domain: in the same release it is rank 9 of 1,224 classified editorial publications across the whole Index, cited in 164 of 15,883 monitored runs (1.03%). It reaches enterprise software answers only when the question is news-shaped. Forbes is the one general publication with real buyer-question reach here, and SaaStr — a trade outlet built for exactly one industry — outranks all four general technology outlets on the news question combined.
How to read the editorial-class rank. The editorial-class rank above is read off the Machine Relations Index's source_role field, which in the current release combines source-type evidence with AuthorityTech's placement catalog; 96 of the 1,231 class members are admitted through the catalog, and the Index is moving editorial classification to source-type evidence alone. The class currently includes six platform domains alongside edited newsrooms: medium.com (class rank 1), amazon.com (7), prnewswire.com (8), apple.com (16), wordpress.com (239) and blogspot.com (330). At the same time substack.com, beehiiv.com and dev.to — the same hosted open-publishing product as Medium — are filed as community platforms. Those six hold 1,358 of the editorial class's 13,525 cited runs, 10.04%, and they include the class's top slot. The class denominator also moves between releases (published figures carry 1,025 to 1,230) because it counts the domains classified in each release. Independent of class membership, the citation rate, cited runs, days cited, engine breadth and confidence on this page are direct per-domain observations over a fixed run set, and are the primary measure; read the rate first. See the MRI methodology and update log.
The practical reading: a SaaS placement in TechCrunch buys attention, recruiting signal and investor legibility. On the evidence in this category it does not buy a slot in the answer a buyer gets when they ask an engine how to choose.
What Your Own Domain Can Win
The share of top-100 slots held by vendor-owned domains is the sharpest split in the category, and it moves with the question.
| Question shape | Vendor-owned in the top 100 | Vendor-owned in the top 10 |
|---|---|---|
| News-driven citations | 34 | 5 |
| Problem-first research | 18 | 4 |
| Comparisons | 12 | 3 |
| Top lists | 10 | 0 |
| Best tools | 9 | 1 |
| Is it worth it | 7 | 2 |
| How buyers choose | 6 | 1 |
Two vendor-owned pages show what earns those slots. Zapier is rank 2 of 258 on problem-first research (12 of 106 runs, 11.32%) and rank 9 of 167 on best tools (11 of 101, 10.89%) — a vendor site that publishes how-to material about a problem class rather than about itself. Netsuite.com is rank 2 of 194 on is it worth it (18 of 107, 16.82%) and rank 6 of 172 on how buyers choose (12 of 99, 12.12%), while sitting at rank 360 of 1,375 on the news question (3 of 596, 0.50%). Both are cited where they answer the buyer's question in their own words, not where they announce things.
That is the owned-source lane for a SaaS company: one page per buyer question shape, written to be quoted, not to convert on the first visit.
How AI Engines Select the Sources They Cite
The engines publish enough about their own retrieval to make the pattern above unsurprising. ChatGPT browses and cites live web sources in its answers (OpenAI). Perplexity documents the crawlers it uses and the conditions under which a site is fetched and cited (Perplexity). Claude's web search tool retrieves and attributes sources at answer time (Anthropic). Google's AI features select and link pages from its index rather than from a media buy (Google Search Central, Google).
Every one of those mechanisms rewards a page that answers the exact question with specifics and a verifiable source, which is what research on generative engine optimization found as well (Aggarwal et al., arXiv 2311.09735). None of them contains a step where general brand prominence substitutes for a page that answers the question.
The buyer-side shift is the reason this matters commercially: B2B buying journeys are dominated by independent, self-directed research rather than vendor contact (Gartner), enterprise AI adoption made AI-mediated research standard practice through 2024 and 2025 (Stanford AI Index 2025), and LLM-referred traffic converts at 30–40%, roughly 10x traditional organic search (VentureBeat, April 2026).
The SaaS AI Visibility Plan, Rebuilt on the Measurement
- Find your category and read its shapes. Enterprise software is one category in the Index; your product may sit in a narrower one. Open the category index and read the leaderboard for each published shape before deciding anything.
- Pick the shapes you can actually win. A company with no news flow should not start at the news question, where LinkedIn leads and 34 of the top 100 are vendor-owned. Problem-first research and comparisons are where a specific, well-sourced page competes fastest.
- Write one owned page per shape, in the buyer's words. The measurement says the slots go to pages that answer a question completely. That is a publishing decision, not a keyword decision.
- Target the publications that hold rows in your category, not the ones on the logo wall. In enterprise software that means Forbes, TechTarget and the category trades — and it means checking the domain profile of any outlet before you pitch it.
- Place where the buyers already are. Reddit is rank 1 on problem-first research and LinkedIn is rank 1 on news-driven citations. Neither is a PR placement; both are participation with a real account and a real answer.
- Re-measure per shape, never as one score. Track your own domain's rank and cited-run count in each shape separately. Six numbers that move independently are the only version of this you can act on.
Frequently Asked Questions
Is a TechCrunch placement worth it for a SaaS company?
For AI visibility in enterprise software buyer answers, the measured answer is no: TechCrunch holds no row on any of the six enterprise software buyer question shapes in this release, and on the news question it is rank 190 of 1,375 sources, cited in 6 of 596 answer runs (1.01%). It remains a strong domain overall — rank 9 of 1,224 classified editorial publications — and it does other jobs well. Buy it for those jobs, not for citation share in buyer answers.
Which publications do AI engines cite for SaaS buying questions?
In the enterprise software category, editorial publications hold 37 of the 600 top-100 slots across the six buyer shapes (6.17%). The ones that recur are Forbes, TechTarget, TechRadar and the category trades. The larger share goes to research databases, implementer and consultant sites, review platforms and the companies' own domains.
Does G2 or Capterra help SaaS AI visibility?
G2 holds rows on five enterprise software shapes, including rank 7 of 167 on best tools (11 of 101 runs, 10.89%), and is rank 9 of 22,213 observed domains across the whole Index. Capterra appears once in the category, at rank 69 of 264 on top lists (3 of 114 runs, 2.63%). Review profiles are worth maintaining; they are not a substitute for an owned page per question shape.
How long does it take to appear in AI answers?
The Index observes each category on its own schedule — the enterprise software buyer shapes are measured over 7 run dates, the news shape over 53 — so movement is visible within weeks of a page being published and retrieved, not within days. The durable predictor is whether a page answers a specific question with sources an engine can resolve.
Should a SaaS company optimize for one AI visibility score?
No. Across the six buyer shapes in enterprise software, no domain leads more than three, and the top-ten composition differs by shape. A single pooled score hides which question you are losing and therefore which page to write next.
Sources and method
Every figure on this page was read at its public Index route on 2026-09-19 by plain HTTP from AuthorityTech's publishing host, release mri_score_v2.0+2026-09-19+0cad03121f60, window 2026-05-10 to 2026-09-19, 15,883 monitored answer runs, 22,213 observed domains, 1,224 classified editorial publications. Category and shape pages read: https://machinerelations.ai/index/categories/enterprise-software and its seven published segments (best_x, how_choose, is_x_worth, problem_first, top_list, x_vs_y, news_topic). Domain profiles read: https://machinerelations.ai/index/domains/techcrunch.com, venturebeat.com, wired.com, businessinsider.com, forbes.com, saastr.com, techtarget.com, gartner.com, g2.com, erpresearch.com, netsuite.com, zapier.com and linkedin.com, each at /index/domains/<domain>. Top-ten and top-100 role counts were computed from the leaderboard rows as published; each leaderboard publishes its top 100 ranked sources. Ranks are rank of the segment's total ranked sources; rates use the segment's own observed-run denominator and are never pooled across shapes. A domain described as holding no row in a shape has no such row on its published profile in this release; the release measures citations in the monitored prompt set and is not a census of everything the engines have ever cited.
See where your SaaS company appears, and does not appear, in AI-generated answers for each buyer question shape. Run your visibility audit →