Skip to content

Application Engineering Intelligence

Understand the knowledge model behind every request: seven categories, fifteen packs, selected per request.

Application Engineering Intelligence, AEI, is Kabaido's working model of the UK industrial supply chain: the knowledge of an applications engineer, built in. It is what lets a request be read in your industry's own terms, with designations decoded, abbreviations expanded and units normalised.

Seven categories, fifteen packs

The knowledge is organised on two levels: seven top-level categories over fifteen knowledge packs. Each pack carries the vocabulary, designation grammars and extraction conventions of one supply area.

CategoryKnowledge packs
Cutting Tools & AbrasivesCutting tools, Abrasives and grinding
Machines, Workholding & AutomationMachine tools and CNC, Workholding and toolholding, Hydraulics and pneumatics, Power transmission
Fabrication & JoiningWelding and joining, Sheet metal and fabrication, Fasteners and fixings
Measurement & QualityMetrology and inspection
Surface Engineering & Heat TreatmentCoatings and heat treatment
Materials & StockMaterials and stock, Plastics and composites
Fluids, Consumables & WorkshopFluids and lubricants, MRO and workshop

Selected per request

Every organisation gets the whole categorisation; there is nothing to switch on or configure. When a request arrives, a prescan of its content selects the relevant packs for that request, so a bearing enquiry is read with power transmission knowledge and a regrind list with cutting tool knowledge. The packs supply the vocabulary; your catalogue supplies the truth.

What AEI does with a request

  1. Reads the text, tables and scanned pages you sent into one document.
  2. Extracts typed line items with normalised units and a citation behind every value.
  3. Matches each line deterministically against your catalogue, your saved designed products and your services.
  4. Assembles the response: matches, clarifications, prices from your own definitions and the action cards.

Trained packs and shells

The fifteen packs are not all at the same depth, and pretending otherwise would be the sort of claim this platform is built to avoid. Some carry a fully worked vocabulary, designation grammar and extraction conventions; others are shells that give the category structure while their depth is still being built. A shell still helps a request be read in the right terms, and it never counts towards escalating a request to a heavier model, because a shell has nothing heavier to say.

What that means in practice is that a cutting tool enquiry and a fastener enquiry are both read, and the cutting tool enquiry is read more sharply. The whole categorisation is available to every organisation either way.

Your catalogue is the authority

The packs supply vocabulary, not truth. They know what a designation means and how the trade abbreviates it; they do not know what you stock, what you charge or what you will substitute. That comes from your own catalogue, your own schema fields and your own pricing rules, which is why the same enquiry read for two different distributors produces two different answers, both correct for the business that received it.

Cite or abstain

AEI never invents a specification, tolerance, price or part number. Every extracted attribute carries a source span, and a value the request does not state stays empty and raises a clarification. Fit scores, match bands and prices are deterministic calculation, not model output; see Working with matches.

AEI powers all five modules. The same knowledge reads an enquiry, drives the Design Engineer on the lines that need designing and sits behind the email agent.