BlogPerspective
Your AI Should Reach Your Systems
The industry spent a decade arguing about data formats and almost none on whether anyone's tools can actually reach each other. Buyers now arrive with an AI client of their own — and the question has changed.
Your AI should be able to reach a supplier's systems and get a real answer — products, pricing, stock, order status — not a PDF. On Brikl, every action in the platform is exposed as a tool an AI client can call, governed by scoped access and one review queue. Brikl for AI, which lets you run the business from your own chat, is coming soon — here is the thinking behind it.
In short:
- Brikl exposes platform actions as tools an AI client can call, so an AI can fetch real product, pricing, stock and order data.
- Brikl re-architected the foundation — roughly twenty backend services consolidated into four — so this was possible at all.
- On Brikl, an AI client authenticates for real, works within scoped tools, and every write it makes lands in the same review queue as every other channel.
- Brikl for AI is coming soon; it ships after v3.0.
For a decade, the promotional products industry has argued about data formats. It has spent almost no time on a simpler question: can anyone's tools actually reach each other? PPAI research finds AI use is becoming ubiquitous among branded-merch firms — which means buyers now arrive with an AI client of their own. The question is no longer "which standard do you support." It is "can my AI reach your system, and what is it allowed to do when it gets there."
Why should your AI reach your systems directly?
A distributor should be able to ask their own AI client for products, pricing, stock and order status — and get a real answer. Today that request usually ends in a spreadsheet emailed back, a PDF, or a phone call. The information exists; it just isn't reachable by a machine acting on your behalf. An open door to the data changes what the AI on your side of the table can actually do for you.
Why can't most platforms do this today?
Because you can't bolt it on. Layering an AI assistant on top of a twenty-year-old architecture doesn't reach far enough — the assistant can only touch what the old system already exposed, which is usually a screen, not the data behind it.
Reaching further goes to the foundation. Brikl consolidated roughly twenty backend services into four so that every action in the interface could be exposed as a tool, cleanly and consistently. That is our engineering choice, not a claim about anyone else — but it is the reason "add AI later" and "build for AI" produce different products.
Who pays for the AI?
The economics flip in the buyer's favor. When a distributor uses their own AI client, that client runs on the distributor's own AI subscription — the token cost is already covered. The platform's job is to provide the door, not the meter. Nobody has to price or resell AI usage; the value is in access, and access is what Brikl provides.
Does open access mean open season?
No. Access without governance is a liability, not a feature. On Brikl, an AI client gets scoped tools — you decide what it can see and do — and it authenticates like any other connection, with no anonymous access. Every write an agent makes lands in the same reviewable queue as a REST call, a webhook or a human edit, with a payload diff, before anything goes live. The door is open; it is also watched.
What does this mean for suppliers?
For a supplier, the work disappears. Each of a supplier's services on Brikl answers over four protocols — SOAP, REST, MCP and PromoStandards — from the same data under the same rules, and Brikl generates, hosts, versions and maintains all of it. There is nothing to build and nothing to run: an AI client connecting over MCP reads the same catalog and places orders against the same state machine as a distributor on PromoStandards.
When does Brikl for AI ship?
Brikl for AI is coming soon — it ships after v3.0. When it does, it will let you run the business from ChatGPT, Claude, Gemini or Grok: every UI action available as a tool, on the same scopes, auth and review queue described here. The principle comes first, and the product follows it. See what's new in v3.0 for what ships now.
Branded merch has spent long enough deciding how to describe a product. The next decade is about who — and what — can reach it. Build the door well, watch who comes through it, and let the buyer's own AI do the rest.
More from the blog
Perspective
The Dashboard Isn't Dying. It's Changing Jobs.
A software founder posted that his users had stopped visiting his dashboard. The replies were more interesting than the observation — and they describe exactly what happens to a promo platform when agents start doing the daily work.
Product
A Store in Ten Seconds: What Turbo Store Changes
When a store took eleven hours to build, only the biggest accounts got one. Turbo Store changes the math — and who gets a store at all.
Product
Brikl v3.0: Bulk Orders and On-Demand Stores in One Place
v3.0 is the release where the two halves of a distributor's business — bulk orders and on-demand stores — finally run on one platform, one catalog, one login.