Discovered Labs AI Visibility Tracker Review 2026

Key takeaways

• This isn't a self-serve SaaS tool. It's an internal instrument Discovered Labs built to run its own AEO/GEO retainers, and you only get view-only access to it if you're a paying client.
• Lacks content generation at scale, AI crawler logs, visitor analytics, and multi-model automation that a platform like Promptwatch offers as self-serve software you control directly.
• The statistical rigor (95% credible intervals, three-run significance checks) is a genuinely good idea that most AEO tools skip, but it's locked behind a four-to-six-figure monthly retainer rather than something you can run yourself.
• Built for B2B SaaS companies roughly in the $2M-$50M ARR range who want an agency to own the strategy end to end, not for teams who want a tool they operate themselves.
• Case study numbers (7x sales-accepted leads, 167% demo growth) are self-reported by the agency with no disclosed sample size or third-party audit.

What it actually is

Discovered Labs is an AEO/GEO agency founded in 2024 by Liam Dunne (who previously scaled Instantly.ai to $20M ARR) and Ben Moore, an ex-Stanford AI researcher who worked on engineering teams at Aurora, Stripe, Coinbase, and Brex. The AI Visibility Tracker is the dashboard they built to power their own client work. It tests buyer prompts across ChatGPT, Claude, Perplexity, Google AI Overviews, and Gemini, and a third-party review also lists Microsoft Copilot as tracked.

Here's the thing that trips people up when they land on this page: it looks and reads like a product you could buy. It has a slick dashboard, FAQ section, feature breakdowns, all the trappings of a SaaS landing page. But the agency is explicit about this in their own FAQ: "We're not a SaaS company. The tracker exists to power the strategy we run for you, so it isn't sold as standalone software." You get it bundled with a retainer, not as a subscription you can cancel and run yourself.

That positioning is actually refreshingly honest, and it changes how you should evaluate this. You're not buying software. You're buying an agency that happens to have built good internal tooling and shows it to you as proof of competence.

Key features

• Prompt cohort tracking. Buyer prompts are sampled from your Google Search Console queries, support tickets, call transcripts, and relevant subreddits, then re-tested every 5 days across every tracked model. The agency says prompts go through coverage, confound, and brand-anchor checks before they're allowed into the rotation, which is a more disciplined prompt-sourcing process than most tools bother to show publicly.
• Statistical significance gating. This is the standout idea here. Every metric ships with a 95% credible interval, and a movement only counts as real if it clears a statistical bar across three consecutive runs. Most AEO trackers, including several well-funded ones, report week-over-week swings without acknowledging that LLMs answer the same question differently every time you ask it. Discovered Labs built their methodology specifically to filter that noise out, and published a blog post calling out competitors for testing in incognito mode rather than logged-in sessions.
• Multi-model coverage. Five engines tracked on the same cadence: ChatGPT, Claude, Perplexity, Google AI Overviews, and Gemini. That's a reasonable spread, though it's narrower than platforms that also track Grok, DeepSeek, Mistral, and AI coding assistants.
• Competitor share-of-voice benchmarking. You configure your competitive set during onboarding (plus whatever the agency surfaces from its own research), and the dashboard tracks share of voice per engine over time. The demo data shows a CRM category example with HubSpot, Salesforce, Pipedrive, and Attio tracked side by side.
• Topical drilldown. You can go from pillar topic down to individual prompt to see exactly where you're winning or losing against named competitors, which turns "we're not visible" into an actual prioritized list of content gaps.
• Citation source mix. A breakdown of which sources LLMs actually cite in your category, split by Reddit, reviews, YouTube, blogs, owned content, and other. This maps directly to the agency's off-page work (Reddit marketing, digital PR, link building), which is the point: the dashboard exists to justify the next line item on your invoice.
• Brand perception scoring. Every LLM response gets scored across dimensions like awareness, competitive positioning, discovery, evaluation, features, integrations, pricing, and reputation, benchmarked against named competitors. This is a genuinely useful idea for catching drift in how AI describes you before it shows up on a buyer's shortlist.
• Radar alerts. Notifications when a competitor publishes a new comparison page, refreshes pricing, or lands in a listicle you're not in. Nice touch, though it reads more like a feature built to fuel urgency in sales calls than a standalone product capability.

Who is it for

This is built for B2B SaaS companies in roughly the $2M to $50M ARR range who want an agency to own their AI search strategy end to end rather than operate a tool themselves. Named clients include Instantly, incident.io, Gladia, Sova Assessment, Granola, Passion.io, Teachable, and Composio, which gives a decent sense of the profile: funded, growth-stage SaaS companies with a marketing budget big enough to absorb a five-figure-a-month retainer and a content team that doesn't have time to run AEO in-house.

It is not a fit for enterprise companies that need complex governance, compliance workflows, or a multi-stakeholder procurement process; a third-party agency roundup (Quoleady) flags this explicitly. It's also not a fit for anyone who wants to own the data and workflow themselves, since there's no standalone purchase path, no public API, and no self-serve trial. If you want a tool you control, this isn't that tool, you'd be buying the agency relationship first and the dashboard second.

Integrations and ecosystem

There isn't much of an ecosystem to speak of because this isn't sold as software. No public API, no Zapier connection, no native Google Search Console or GA4 integration documented anywhere. Slack is used for client communication (weekly async updates, bi-weekly live calls), not as a product integration. There's no browser extension or mobile app. If integrations and a developer-friendly API matter to you, this isn't where you'll find them.

Pricing and value

Pricing is structured around agency retainers, not software tiers:

• Search Visibility Diagnostic: €4,370 one-off, the only low-commitment entry point
• Establish retainer: €6,995/month month-to-month, discounted to roughly €5,995/month effective on a 6-month commitment. Includes up to 20 content units per month, a one-off technical SEO audit, brand consistency audit, landing-page optimization on up to 2 pages/month, and Reddit engagement scaled to 10 mentions/month
• Compete retainer (the agency's "most chosen" tier): €10,995/month, discounted to roughly €8,995/month effective on a 6-month commitment
• Outlier and custom projects: from €15,995/month
• A standalone Reddit marketing add-on for teams handling content in-house: from €4,995/month

There's no free tier, because there's no product to subscribe to. Compared to self-serve AI visibility platforms, where entry pricing often starts under $100/month, this is a different category of spend entirely. You're paying for a dedicated team (SEO manager, off-page specialist, content editor, plus AI/ML engineering support), not software access. Whether that's good value depends entirely on whether you want to do the work yourself or pay someone else to do it, which is really the core decision this whole page is trying to get you to make.

Strengths and limitations

The statistical rigor is the clearest strength here. Gating movement claims behind credible intervals and requiring three consecutive significant runs is a smarter approach to AI-answer volatility than most tools bother with, and it's backed by a published methodology post rather than just marketing copy. The topical drilldown and citation source mix are also genuinely useful for turning visibility data into a content plan, which is the whole point of the Discovered Labs pitch.

But as a tool, it has real gaps next to a dedicated AI-visibility platform like Promptwatch. There's no AI crawler log analysis showing when ChatGPTBot or PerplexityBot actually hits your pages, no visitor analytics tying AI citations to real site traffic and conversions, no automated content generation or CMS publishing pipeline, and no Agent Chat or Unified Actions workflow for turning insights into prioritized tasks on your own. It tracks five engines versus the 10+ covered by broader platforms, and there's no self-serve way to adjust prompts, competitors, or cadence without going through the agency. The case study results (7x sales-accepted leads at Gladia, 167% demo growth at Sova) also come with no disclosed sample size or independent verification, which is worth keeping in mind when evaluating the pitch.

Bottom line

If you want an agency to own your AI search strategy and you're comfortable paying €6,995 to €10,995+ per month for it, Discovered Labs' tracker is a reasonably sophisticated instrument backing that service, with better statistical discipline than most competitors show. But if you want a tool you can run yourself, adjust on your own schedule, and use to actually generate and publish optimized content without going through an agency, Promptwatch is the stronger fit: it covers more models, tracks AI crawler behavior and real traffic attribution, and includes Content Agents that plan, write, and publish GEO content directly to your CMS, which the Discovered Labs tracker simply doesn't offer as a self-serve capability.

Promptwatch

Track and optimize your brand's visibility in AI search engines

View more