An agency asked me to build them a brand and social outreach tool last year. They wanted something that helped the team identify the right brands to approach for each athlete they represent, track every partnership conversation in motion and time the outreach so it did not compete with itself.
A reasonable thing to want.
I asked how they were tracking those things today. The answer, more or less, was: email, Slack, a shared Google Drive, a spreadsheet one partner maintained and several other spreadsheets nobody quite knew about.
I told them I would not build them the tool they had asked for.
This was not because the idea was bad. It was because I have watched fifteen people try to maintain data entry into an outreach tool and I have watched zero of them succeed past month three. The data-entry burden was the problem and a new tool would be the same problem with a fresh UI.
So we built the opposite.
The system reads the senior team’s inboxes, reads the shared drive, reads the meeting-notes channel, reads the documents that already exist. Every night. It extracts the athletes, the brands, the active conversations, the relationships, the signals. It writes those into a structured database. The humans do not have to enter a single thing they did not already have to type somewhere else.
The partners came in on Monday and the knowledge base was full. Not because they had spent two weeks doing data entry; they had done no data entry at all. Because the data was already there, distributed across twenty thousand emails and a few hundred documents. The system read it faster than they ever could have.
The outreach engine built on top of that knowledge base could then do its job: suggest which brands to approach for which athletes, flag where a conversation had gone quiet, surface the signals from news and email that meant an opportunity had opened. The social outreach layer worked the same way, reading what the team had already written and drafting what came next.
The pattern generalises beyond outreach. Every business has, right now, a huge amount of structured knowledge sitting in unstructured form inside systems nobody thinks of as knowledge bases. The email archive. The contract folder. The supplier folder. The meeting-notes channel. Over time, the team’s institutional knowledge lives in there.
Most software vendors think the solution is to give the team a better place to write things down. This is the wrong answer. The team already wrote things down, years ago, in the first system. The solution is to read what they wrote.
I would say this to any business considering an outreach tool, a CRM, a knowledge-management system, an internal wiki or a decision log. Before you make your team type anything new, audit what they already type. There will be more of it than you think and an AI that reads it is cheaper than a team that keeps writing it twice.
The second database costs less than the first one, because the first one already exists.