Back to Insights
AI Readiness|6 min read

The Failure You Cannot See

You can spend a year enriching product data for a catalog that no machine can actually reach.

By Ashley Stennis, Director of Growth, Luminos Labs

Read anything about AI readiness in B2B commerce, and you get the same advice. Clean your product data. Enrich it. Get the taxonomy right. I don't disagree with any of it; we do that work for clients, and it pays off.

All of it assumes something nobody checks, though: that a machine can reach your catalog at all. For a lot of distributors, it can't. Everything sits behind a login, so there is nothing on the open web to read, and you can spend a year enriching the contents of a locked room.

Key Takeaways

  • Enriching product data only matters if a machine can reach it. Many B2B catalogs sit entirely behind a login, invisible to the AI tools buyers now use to build a shortlist.
  • The gate made sense for twenty years: pricing is relational and needs protecting, so putting a wall in front of everything was the cheapest thing to build.
  • That trade-off is now costing distributors deals they never see, because a lost evaluation leaves no trace in the CRM the way a lost deal does.
  • One distributor moved from ~24 self-registered accounts a month to ~260 simply by removing the requirement that a rep manually open every account.
  • The fix is separating two things the gate bundles together: contract pricing (which must stay protected) and product existence/fit/availability (which doesn't).

Why the door is locked

Nobody did this by accident. B2B put its catalog behind a login for reasons that made sense then and mostly still do.

Pricing here is relational. What a customer pays comes out of their contract, their volume, their buying group, how long they've been with you. You obviously can't publish that. And once you accept it, putting a wall in front of everything is far and away the cheapest thing to build, because per-account visibility rules get complicated fast, while a gate settles the whole question in one decision. There's also the older instinct every distributor inherits somewhere along the way. Don't let the company across town see your book.

So the gates went up, and for twenty years they cost almost nothing. Buyers found you through a rep, or a catalog that showed up in the mail, or somebody they met at a trade show. Search was where consumer brands fought each other. I spent a good part of my career on the other side of that, telling people they were invisible. Said to a distributor, it was advice they were mostly right to shrug off. The buyers who mattered were not looking there anyway.

That last part has stopped being true, and I don't think many people have noticed yet.

You will never catch it happening

Buyers with a specification to fill used to open five supplier sites and compare them. Plenty of them now put that specification to an AI tool and work from whatever comes back. The shortlist they end up with is built out of pages the model was able to read. If your specs and availability sit behind a login, you didn't lose that comparison. You were never in it.

If your specs and availability sit behind a login, you didn't lose that comparison. You were never in it.

Which is what makes this so hard to root out. Lose a deal, and at least you have something to work with, somebody in your pipeline who went elsewhere and can be asked why. An evaluation you were never part of leaves nothing behind at all. Nothing in the CRM, no quote that came back, nobody complaining. Only a number that lands a little lighter than it should, quarter after quarter, and no one can tell you the reason.

So it survives in companies that are otherwise very well run. Nobody is ignoring a warning sign. There isn't one. You find this by deciding to go and look for it.

What happened when one of them opened up

A distributor we work with carries around 60,000 items from 600-odd brands and does roughly $250 million a year in sales. All of it behind the gate. Their order path required a person at nearly every step, too, so customers couldn't register themselves, and a rep would call back to finish the order.

New customers came in at maybe two dozen a month, every account opened by hand. Seventeen years of that. Once the gate came down and registration went to self-service, they moved to around 260 registrations a month, of which about 230 were approved.

All that time, two dozen a month was just the number. Nobody there thought of it as low, because there was nothing to hold it against. That is the shape of this problem from the inside. The ceiling reads as the market, right up until somebody lifts it.

What you sell is not what you charge

This is the part I'd want a technical leader to take away.

The gate bundles together two things that don't belong together. Contract pricing has to stay protected, and it can. But the fact that a product exists, what it does, what it fits, whether you have any, none of that needs protecting.

Most distributors are hiding the second thing to protect the first, because ten years ago, that was the cheap way to build it, and nobody has looked at the decision since.

It's fixable. How big a job that is depends on how tightly your pricing logic is wound into your catalog, and that varies enormously. For some distributors it is a configuration change. For others it means untangling twenty years of assumptions about who is allowed to see what. Either way, you can find out which one you are in about a week.

The transaction side works the same way. As more buying gets handled by software acting for the purchaser, an order path that waits on a rep to call back for card details isn't slow. It's closed. Anything in that path requiring a human to be available will fail at ten at night, and will increasingly fail when whatever's at the other end was never human to begin with.

Contract pricing has to stay protected, and it can. But the fact that a product exists, what it does, what it fits, none of that needs protecting.

Two things to go and check

Search a part number you sell. See whether your own site comes back.

Then walk your order path and ask what genuinely needs a person standing there, as opposed to what's simply nicer with one.

Neither takes a budget. Both tell you whether a budget would change anything. None of it is clever work; it is dull, which is usually why nobody has done it.

Somewhere in your business is a number you have come to think of as your market. It might be exactly that. It might also be the size of the door you left open. Nothing is going to tell you which, because this failure never announces itself. It waits for somebody to get curious enough to check.

Ashley Stennis

Ashley Stennis

Director of Growth, Luminos Labs

Curious whether your own catalog is reachable? It's a quick, no-budget check, and we're happy to walk through it with you.