Glean is a genuinely strong product, built for a specific kind of organization: large enough to have real enterprise search needs across a hundred connected systems, with the IT resources and budget to match a real procurement process. For a twenty-person agency managing a dozen active clients, that's a mismatch on almost every dimension that matters, not because Glean is a bad product, but because it's solving a differently shaped problem.
Why the size mismatch is the real issue
Enterprise search tools are built around the assumption of scale: many departments, many systems, a dedicated IT function to manage the rollout. A small agency doesn't have most of that. The actual problem a twenty-person agency faces isn't "we have too much documented knowledge spread across too many systems to search efficiently." It's closer to the opposite: most of what matters about each client relationship was never formally documented anywhere, it lived in calls, in Slack threads, in decisions made quickly and never written up. Enterprise search, however good, can't surface knowledge that was never captured as a searchable document in the first place.
What a small agency's actual problem looks like
The real daily friction at agency scale is closer to what's described across multiple accounts of agency context switching: an account manager moving between client accounts throughout the day, needing accurate, current context for whichever client just came up, without a heavyweight system to manage or IT overhead to maintain. That's a fundamentally lighter-weight need than what enterprise search platforms are built to solve, and paying for enterprise-scale infrastructure to solve it is likely to mean paying for capability that never actually gets used.
What "lightweight" should actually mean here
A memory layer built for agencies this size should work from day one, without a lengthy implementation project, connecting to the handful of everyday tools a small agency actually uses, Slack, email, whatever project tool is already in place, and capturing client context as it's naturally created rather than requiring anyone to formally document it first. It should also handle the specific failure mode small agencies hit constantly: keeping each client's context cleanly separated from every other, so nothing from one account ever bleeds into work being done for another.
The honest comparison
Glean, at agency scale, is solving a problem you likely don't fully have yet, with a cost and complexity built for an organization you likely aren't yet either. A lightweight, agency-sized memory layer solves the problem a small team actually has: too many client relationships for any one person to hold accurately in their head, none of it formally documented anywhere Glean-style search could find it. Worth picking based on which problem is actually yours, rather than which product has the stronger enterprise reputation.
If the lightweight version sounds closer to your actual situation, it's worth a look at indexbrain.online.
