A solo founder in Cambridge has built a tool that promises to solve a problem most software companies barely know they have—whether ChatGPT, Claude, and other AI assistants are recommending their products when users ask buying questions.
CiteGraph, launched this fall by a developer who goes by Asif or @akbuilds_ on X, tracks brand mentions and citations across four major AI engines, then automatically generates content designed to win those recommendations. The service starts at $29 a month, undercutting some competitors in what some are calling "answer engine optimization," a still-forming market.
The pitch hinges on data Asif's team has collected: the company states it has measured more than 81,000 citations (a figure posted on the homepage in mid-October but not independently verified), and CiteGraph found that nearly half of all sources leading AI assistants cite when recommending products come from third-party listicles rather than vendors' own websites. Only 36 of 70 top-ranked products in a seven-category sample were sourced from their own domains, according to research the company published in late August.
"Most founders think if they rank on Google, they're visible," Asif wrote on PeerPush in September, where CiteGraph briefly held the top Product of the Day slot. "But ChatGPT and Claude pull from comparison sites, not your homepage."
Measuring what AI sees
The platform runs 10 buyer-intent questions per project through all four engines—ChatGPT, Claude, Gemini, and Perplexity—three to five times each, capturing verbatim answers and the URLs cited as sources. CiteGraph then calculates confidence intervals on two distinct metrics: whether your brand gets mentioned at all, and whether your own domain appears as a cited source.
The distinction matters. A startup might be named in an AI-generated answer but sourced from a competitor's comparison page or a software directory instead of its own site, diluting brand control and losing the click. CiteGraph flags these "lost answers" where rivals appear but you don't, then surfaces a prioritized list of which questions to tackle first.
In a blog post published in September and updated early the following month, the company reported examining 45,309 citations across 7,260 AI-generated answers in a specific research sample. Listicles, roundups, and competitor product pages accounted for 91.9 percent of citations in that sample—a finding that suggests AI assistants rely heavily on aggregator content rather than primary sources.
Auto-generated fixes, with footnotes
Where CiteGraph diverges from simple monitoring tools is its content engine. The platform's "Write the page" feature ingests up to 40 pages from a customer's existing site to build what Asif calls a "fact and style ledger," then drafts comparison articles, FAQ pages, and buying guides with every sentence linked to a source. Drafted pages export in Markdown or HTML with structured FAQ schema, ready to publish.
An optional Autopilot mode goes further, selecting topics based on lost answers, cited competitor pages, and Google autocomplete suggestions, then writing and committing posts directly to a customer's GitHub repository in the site's existing design. The system includes verification checks to ensure numbers and links trace back to sources, and it monitors whether AI crawlers have indexed the newly published pages.
The tool re-measures citation rates in the next weekly scan to track movement, though the company doesn't disclose how long it typically takes for new content to register with AI models whose training data and update cycles remain opaque.
Pricing against a crowded field
CiteGraph's Starter plan runs $29 monthly for two domains, 10 questions, three runs per question across four engines, 30 drafted pages, and one seat. Growth climbs to $69 for three domains, five runs, 60 pages, and five seats. Scale costs $149 for six domains, five runs, 200 pages, and unlimited seats. All tiers include a three-day free trial.
The company backs its service with a 120-day guarantee: if a customer's brand isn't named by an AI assistant within four months, CiteGraph refunds all subscription payments—provided the customer kept weekly scans running and published at least three drafted pages.
That pricing undercuts Profound, which positions itself as a full answer engine optimization platform and starts at $99 monthly billed annually for visibility dashboards and citation rank tracking. Otterly.AI offers AI search monitoring at the same $29 entry point as CiteGraph, while Scrunch AI markets an "Agent Experience Platform" with FAQ pages that explicitly cover citation tracking.
The competitive landscape is still forming. Answer engine optimization lacks the mature tooling and consensus practices of traditional SEO, and it's unclear whether winning citations in today's AI models translates to revenue or simply shifts the channel where brands compete for attention.
Early traction, cautious claims
Yusuke, founder of LandingBoost, wrote on X in early September that he "tried CiteGraph by AK and immediately found the exact AI visibility gap LandingBoost needed to fix." He turned the insight into what he called a data-backed page using 48 SaaS scans. A PeerPush reviewer described the tool as "brilliant for tracking competitive landscape," noting that the AI analysis feature saved manual research time.
Asif said the platform includes an in-product assistant called Ask, powered by Claude via OpenRouter, that can draft pages, flag recommended moves, and re-run scans. The assistant is capped at 150 messages daily per user.
One wrinkle in CiteGraph's published research: some of the company's data carries timestamps dated to late August 2026, which appears to be an error or placeholder text. Asif did not immediately respond to questions about the discrepancy or whether the cited studies reflect actual historical data. That kind of detail matters in a market where credibility hinges on measurement rigor and where vendors are selling the promise of visibility in systems whose inner workings they don't control.
For now, CiteGraph represents a bet that the next battleground for software discoverability isn't Google's blue links but the synthesized answers served up by conversational AI. Whether that shift is happening fast enough to justify a new category of tooling remains an open question.
