The program was delivered by Discovered Labs.
Sementa was not the delivery agency. This is our independent analysis of the agency’s published results, reviewed on October 1, 2026.
The challenge
Make discovery lead to product use.
Granola, an AI notepad, needed to reach buyers across Google and AI search in a competitive category. The reported program combined content, technical SEO, links, and community work, measuring Google and AI search separately.
The useful question for a growth team is what happens after discovery: do the people who arrive understand the product, try it, and reach the moment that makes it valuable?
The published results
Three metrics. Different jobs.
47×
Monthly new active users attributed to AI
Compared with January 2026
Reported as +4,600% growth, reaching thousands per month. The source does not specify an end month for this metric.
+558%
Non-branded AI citation rate
January–August 2026
Relative growth measured in Discovered Labs’ AEO platform; not a 558 percentage-point increase.
+739%
Non-branded AI mention rate
January–August 2026
Relative growth in brand mentions, tracked separately from website citations.
These are provider-reported figures. The public page does not supply a full prompt sample, engine breakdown, or attribution specification. A combined program cannot isolate the effect of any one tactic.
What the program included
Content, access, and relevant sources.
Content
New and improved pages for product features, audiences, and commercial questions.
Technical work
SEO checks and implementation, with attention to crawlability and important pages.
External sources
Relevant Reddit communities and link building alongside the website work.
Sementa’s interpretation
What we would carry into the next brief.
These are our recommendations for applying the lessons to a new project. The starting point is your own buyers, website, and measurement baseline.
01
Choose the business event before the visibility metric.
A mention can put a product on a shortlist. A citation can send a visitor to a page. Activation tells you whether that visitor reached a useful product experience. Treat these as separate stages, each with its own denominator.
Put it into practice
For an app, define an activation event with your product team. For a sales-led SaaS company, agree what makes an inquiry qualified. Then connect identifiable AI referrals to that event where the data allows.
02
Give each buying question a useful destination.
A page should help someone make a decision: whether the product fits their workflow, what it integrates with, or how it compares with alternatives. Adding pages only helps when each one answers a real question with verifiable detail.
Put it into practice
Map important questions to existing URLs. Improve a weak answer, commission a missing one, or collect product evidence first. Assign a purpose and a reviewer to every content brief.
03
Review the sources around your website.
Your own pages are only part of a buyer’s research. Relevant product directories, reviews, and community discussions can add useful context or expose information that needs correcting.
Put it into practice
Record the sources observed in your tracked answers. Check the facts and update profiles you control; make useful, transparent contributions where participation is appropriate.
04
Keep the comparison stable enough to interpret.
An impressive percentage needs a baseline. Changing the questions, engines, market, or tracking method can change a visibility score even when the underlying pages stay the same.
Put it into practice
Keep a dated question set, record engines and run conditions, and retain answer evidence. When scope changes, establish a new baseline. Review qualified inquiries alongside mentions and citations.
