AI Search as Retention Strategy

This reveals that AI search is a retention feature first, not a profit center. Ecosia still makes its money the old fashioned way, from ad bearing searches, so paying Exa around $300,000 per month only makes sense if AI overviews stop users from defecting to Google, Microsoft, and Perplexity. In practice, Ecosia shows AI on roughly 30% to 40% of queries, keeps plain search on about 70%, and measures success by whether those users come back and later do monetizable searches.

• The workflow is simple. A user asks a harder question, Ecosia routes it to Exa, Exa returns a summarized answer with links, and Ecosia puts that above normal results. That answer costs Ecosia money on every triggered query, while a normal navigational or shopping search can show ads immediately.
  1 sacra
• The retention lift is concrete. Ecosia A.B. tests users who see AI overviews versus users who do not, and reports a 5% to 10% lift in 7 day retention for new users, plus a 3% to 5% lift at 30 days. That makes AI search look more like paid acquisition or product insurance than direct monetization.
  1 sacra
• This is the same monetization split now shaping the whole category. Perplexity has had to layer in subscriptions, premium ads, and commerce because AI answers alone are expensive, while Exa sells the picks and shovels through usage based pricing that passes those query costs on to customers like Ecosia.
  2 sacra 3 sacra

Going forward, consumer search products will use AI answers selectively, on the queries most likely to improve habit and brand perception, while preserving classic ad inventory on the bulk of searches. That pushes search toward a hybrid model where AI wins attention, but plain search still pays the bills, and infrastructure vendors like Exa capture value each time that handoff happens.