Get 25 battle-tested prompts
Turn real client conversations into Linkedin content that builds authority and attracts the right clients.
Turn real client conversations into Linkedin content that builds authority and attracts the right clients.
Last week, a founder I know through a community was offered $500,000 for his agency's internal operations documentation.
Not for the agency itself. Not for the client list, the team, or the recurring revenue. Just the documentation. The process docs, the playbooks, the internal "here's how we actually do this" material that nobody outside the company has ever read.
He turned it down.
Most founders would have laughed it off as a joke. He didn't, because he understood exactly what was being bought: agency institutional knowledge, which has quietly become one of the most sought-after assets in the AI economy.
In August, Google was named the winning bidder at a bankruptcy auction for Spirit Airlines' internal records at $10 million, covering emails, contracts, and the operational history of a company that no longer flies. The data on offer included roughly 100 million emails and 500 million Microsoft Teams chats, plus spreadsheets, financial databases, audits, and project management documents.
Worth noting: the deal isn't done. The sale is still awaiting court approval, with a hearing set for November. The privacy ombudsman only backed it after the scope was narrowed to exclude consumer transaction, refund, inflight and WiFi systems, and unions and pilot groups have objected. That fight is itself the tell. Nobody litigates this hard over data nobody wants.
Google wasn't alone at the auction. AI data company Mercor opened at $7.5 million. Another firm, micro1, has been publicly offering businesses up to $2 million for their internal operational data, including CRM records, SOPs, and communication history. Data brokers sourcing this material advertise ranges from $100,000 to $4 million depending on volume and depth.
This is not a one-off. The Washington Post's review of recent securities and court filings describes an AI licensing spree running from tens of thousands of dollars to a quarter billion. It is a systematic effort to acquire something the models have never had access to.
Agents are information-rich and context-poor.
Every major model has been trained on virtually everything that has ever been published publicly. Textbooks, forums, case studies, Wikipedia, GitHub, the entire indexed web. The general knowledge problem is largely solved.
What agents have never seen is how a real business actually operates. The order of operations for onboarding a new client. The decision an experienced account manager makes at 11am on a Tuesday when something goes sideways. The three questions your senior consultant asks before committing to a recommendation. Step one through ten, with the outcome attached, run thousands of times across a decade of real engagements.
That is not on the internet. It has never been on the internet. And labs understand that building agents capable of doing real knowledge work requires exactly this kind of data.
One founder I spoke with put it plainly: when a key employee went on vacation for ten days, the team's response was "this should be great because now we can test what would that person do." The expertise was completely inaccessible outside of that one person's head.
That is the gap the labs are trying to close. And they are paying for it.
For a long time, gating expertise behind retainers and hourly engagements was the business model. You sold access to the thinking, one hour at a time. The knowledge itself was protected not because it was codified and secured, but because it lived exclusively in people's heads.
That felt like protection. In practice, it was fragility.
Across hundreds of conversations with agency founders, the pattern is consistent. The most valuable knowledge in the firm is almost always the least accessible. It walks out the door when someone leaves. It becomes inaccessible when a senior person goes on vacation. It sits in old Slack threads, outdated deliverables, and the institutional memory of whoever has been there longest.
One agency owner described his firm's internal knowledge management as a "Zettelkasten graveyard." Another said their most experienced person had 570,000 words of recorded thinking that the rest of the team could not access. A third was spending 100 hours a month on manual reporting because there was no way to surface what the firm already knew.
The expertise was real. The problem was that it had never been made portable.
What clients actually pay for is not the execution. It is the judgment baked in from years of seeing the same problems across dozens of different companies. The consultant who has solved a problem 47 times sees things a first-timer cannot. That pattern recognition, that calibrated instinct, that is what commands the premium.
The firms that understand this are starting to treat it accordingly. They are documenting decision trees, not just deliverables. They are capturing the thinking behind the output, not just the output itself. They are building internal systems that make the firm's accumulated judgment queryable, not just the work product.
One fractional consultant I spoke with built a diagnostic tool that encodes years of client assessments into a structured intake process. The tool identifies revenue leaks and strategic gaps based on patterns he has seen across hundreds of engagements. What used to take a 90-minute intake call now surfaces in minutes, and the output is more consistent than any individual's recall on a given day.
That is not just an efficiency gain. It is the firm's IP becoming portable.
The window for treating proprietary knowledge as an informal competitive advantage is closing. It is closing because the firms that codify it will start compounding on it, internally and in how they deliver to clients. And it is closing because the data market for operational knowledge is now real, which means the value of that asset is being recognized externally even if it isn't internally.
The question is not whether your firm's knowledge is valuable. If you have been in business long enough to have developed real patterns, it is. The question is whether it lives in your systems or in someone's head.
One agency owner I spoke with described their digital director's custom internal repository as a tool that "speeds up his own work but can't be accessed by the rest of the team." That is the state of most service firms right now. The IP exists. It is just siloed.
The firms that move in the next 12 to 18 months will have a meaningful head start. Not because they built something complicated, but because they made a decision to treat what they know as an asset worth protecting.
None of this requires a major technology investment to begin. The firms making real progress started with a simple audit: if your two most senior people left next month, how much of what clients pay you for walks out with them?
If the answer is "most of it," the documentation isn't the problem. The structure is. The knowledge exists. It just hasn't been captured in a form that survives beyond the person who holds it.
That is the starting point. Not a new platform or a new process. Just the honest acknowledgment that the firm's most valuable asset has no form yet.
The labs bidding on Spirit Airlines' emails and your acquaintance's playbooks for millions of dollars already know what it's worth.
The question is whether you do.
