Your local knowledge is a sales asset.
The best local knowledge lives entirely in conversation, asked for and answered in person. That's exactly the knowledge an AI agent can't reach, however good the website is.
4 min read
Why this knowledge is worth more than it looks
Say a small guesthouse sits a short walk from a beach that floods its car park at the highest spring tides, a detail every member of staff mentions to anyone driving. It's never been written down, because it's never needed to be: a guest checking in gets told, a phone enquiry gets told, the knowledge moves person to person and has always been enough. An AI agent recommending the guesthouse to someone planning to drive has no one to tell it, and no way to know the detail exists at all.
This is the pattern behind most of what a good local business knows that a website doesn't say: which room gets the view, what a first-time visitor always asks and what the honest answer is, which nearby path floods, which local event fills the village up for a weekend. None of it is secret. All of it lives in conversation, repeated so often it stopped feeling like information worth writing down.
A generic description of a coastal guesthouse could be written by someone who's never visited. The detail about the car park at spring tides could only come from someone who runs the place, which is exactly why it's valuable: it's the kind of fact that makes a recommendation feel like local advice rather than a listing. An AI that can pass on that detail is doing something a generic search result never could. An AI that can't has nothing to distinguish this guesthouse from any other one near the same beach.
What's worth writing down, and what isn't
Not everything staff know needs recording, and trying to capture all of it produces a document nobody reads and nobody keeps current. The test worth applying: would a guest be glad to have known this in advance, and would they have had no way to find it out themselves before arriving? The spring tides detail passes easily. A staff member's personal opinion on the best table doesn't, unless it's genuinely what most guests ask for. Specific, practical, and otherwise undiscoverable is the bar, not comprehensive.
A useful second question, once the first has narrowed the list: has this come up more than once? A detail one guest asked about a year ago probably isn't worth writing down. A detail every member of staff has repeated dozens of times, without anyone deciding to, is exactly the kind of thing that's become common knowledge internally while staying invisible to anyone who hasn't stayed there yet.
Writing it down is only half of it
A sentence on a website is still just a sentence, the same gap this hub keeps coming back to. "The car park floods at spring tides" is more useful written down than left in conversation, but it's still an unsourced claim until it says who confirmed it and when: staff-verified, checked against the tide table each season, rather than a fact with no one standing behind it. The difference matters here for the same reason it matters everywhere else. A written claim with no source is only marginally more checkable than an unwritten one; it's just easier to find.
Turning it into something an AI can use
Once a piece of local knowledge is written down plainly, with a source, it stops being a fact one guest gets told by luck and starts being a fact any AI recommending the guesthouse can check and pass on with real confidence to every guest who needs it. That's a small amount of writing for a large amount of reach: the same sentence that took a member of staff thirty seconds to say once now answers the same question for everyone who asks an AI instead of asking at check-in. That's the specific gap Selfe exists to close: turning what staff already know and already say, out loud, into sourced facts an AI can check and repeat with the same confidence a local would.
How do we find out what we know that isn't written down?
Ask whoever's on the front desk what they tell guests every week without being asked. That list is almost always the right starting point.
Won't this take a long time to capture properly?
Usually not. Most businesses have a handful of these facts, not hundreds, and each one takes a sentence or two once someone thinks to write it down.
The first thing an agent gets wrong about your business.
An AI agent rarely gets a business wrong out of carelessness. It gets it wrong because the fact that would have changed the answer was never written down anywhere it could check.
Structured data, explained without the jargon.
A returns policy that says "just get in touch if there's a problem" and one that says "returns accepted within 30 days, unworn, tags attached" can describe the same shop. Only one of them can be checked by something that isn't a person using judgement.
The things an agent cannot guess.
Getting found and getting matched are two different jobs. A business can nail the first and still lose the second, because an AI agent only knows what it's been told, never what it could have worked out by being there.