How Granola grew to thousands of monthly activated users from AI search

A complete SEO and AEO plan needs more than a keyword export. Discovered Labs builds one in five steps:

1. Define the market from the product, customers, sales conversations and existing content.
2. Map buyer decisions into Google search keywords.
3. Map those decisions into AI prompts covering scenarios, constraints and comparisons.
4. Measure Google visibility and track when AI answers mention the brand or cite its pages.
5. Turn each insight into an action: create new content, refresh an existing page, fix a technical issue or earn coverage from a credible external source.

The company and its buyers define the market. Search tools validate and expand it.

During the engagement, community work focused on Reddit. The team also expanded feature- and persona-led content, optimized an AI note-taking page and focused commercial topics on questions buyers ask when comparing options.

Most content plans use two familiar signals: keyword gaps for new pages and falling rankings for refreshes. Both matter, but neither shows whether the content covers the scenarios, constraints and comparisons buyers bring to AI, or whether a weak page should be overhauled, improved or left alone.

Discovered Labs uses five decisions:

1. Map each priority search intent to an existing page or a real content gap. Keep weak pages separate from missing pages.
2. Create new content when no focused page can do the job.
3. For existing pages, assess rankings and SEO performance separately from the page's usefulness as a source for AI answers. Discovered Labs' CITABLE framework checks whether the content is clear, grounded, structured and current.
4. Choose one action: preserve what works, improve SEO, optimize for AI answers, overhaul the page or collect more evidence.
5. Brief the exact change, publish it and retest Google and AI visibility separately.

This avoids creating a page for every keyword variation or broadly rewriting content that already works.

The team created feature- and persona-led content for Granola in May, optimized an AI notepad page in June and focused on more commercial topics in July.

Technical SEO audits often end as color-coded spreadsheets. Every red or amber issue looks urgent, the client inherits the implementation work, and a higher site health score becomes the definition of success. Tool severity alone cannot show which issues affect pages that matter to the business.

Discovered Labs uses three steps:

1. Prioritize issues by business impact. Validate each issue, tie it to an affected page and rank performance, crawlability and indexation problems by their likely effect on discoverability.
2. Implement the fixes that matter. Search and engineering work together on high-impact changes. The client handles only changes that genuinely require its access or approval.
3. Verify the fix. Recheck the original performance, crawlability or indexation issue. Where relevant, also check whether answer engines can access the page.

A technical audit becomes useful once the team knows what matters, makes the change and checks the affected page.

Off-page SEO often starts with authority scores and backlink targets. Authority matters, but those inputs do not show whether a source is relevant to a buyer question, describes the company accurately or offers a credible way to act.

Discovered Labs uses four steps:

1. Find the exact source tied to a priority buyer question, such as a page, thread, review or directory listing.
2. Check what it says. Is the company absent, inaccurate, outdated, incomplete or already represented well?
3. Choose a credible route. Discovered Labs' Own / Claim / Earn / Contest (OCEC) framework separates sources the team controls, profiles it can claim, coverage it must earn and inaccurate claims it may need to contest.
4. Take the action and record what changed: update owned information, claim a profile, contribute something useful, earn a relevant link or leave the source alone.

Each link or mention has one job across Google and AI search: improve relevant coverage or keep company facts consistent.

For Granola, the team moved Reddit work from broad B2B communities to communities closer to the product in April, then began link building in June.