AI for industrial sales
Applied Intelligence
Application engineering across the eight industrial domains. Kabaido reads RFQs and POs the way an industrial sales team would: every value typed, normalised and cited to the customer's own words before any matching begins.
Can you quote the following please, needed by 26th June: 10 off 10mm 4FL carbide endmill AlTiN coated, LOC 30 min
RFQ-2026-aero.pdf- L110×10 mm 4 flute carbide endmill, AlTiN, LOC 30 mm min
- L25×6 mm 2 flute ball nose endmill, uncoated
- L320×6.35 mm (1/4 in) 3 flute square endmill, TiAlN
Every value cited. Missing details become questions, never guesses.
+1 more lineEnvelope first
Structured first, whatever arrives
Whatever form the request takes, it becomes one structured envelope before anything else happens. Envelope first and always.
- Pasted text
- An email body or a handful of lines pasted straight in. The structured envelope comes first, whatever the formatting.
- An RFQ or purchase order arriving as a document is read into the same envelope, page by page.
- Spreadsheet
- A line list arriving as a spreadsheet becomes typed request lines with normalised units.
- Image page via vision
- A photographed or scanned page is read by vision. Every extracted value still carries its source span.
A forwarding email address per organisation is on the roadmap.
Application context matters
A match is not a keyword hit. Kabaido reads the application context and turns it into a fit rating you can audit field by field, so a recommendation can be checked rather than trusted.
- Workpiece material
- Read here as 316 stainless. A coating or substrate suited to stainless scores differently from one ground for aluminium.
- Operation
- Slotting, profiling or finishing changes which geometry fits. When the request does not state the operation, the field stays open and is asked, not assumed.
- Machine
- Read here as BT40 VMC. Taper and spindle bound the holder and the shank that can run on it.
- Tolerance
- A stated tolerance narrows the candidates that can hold it. Left unstated, it becomes a clarification rather than a silent default.
The same fields carry through to matching, costing and the quote. See how a matched line is priced on Configure.
Cite or abstain
Every value points back to the text that produced it
Every extracted attribute carries a source span, so a quoted value can be checked against the customer's own words. When the request does not state a value, the envelope carries a clarification instead of a guess.
Cite or abstain
The knowledge core
Four industries, eight knowledge packs
Customers self identify by what they make: sawing, machining, cutting or grinding. Beneath the four sit eight knowledge packs, each carrying the attribute dictionaries, abbreviation lexicons and unit conventions the work leans on.
- SawingCircular and band saw blades, from the request to a quoted blade.
- MachiningCNC cutting tools: end mills, drills, inserts, reamers and taps.
- CuttingIndustrial machine knives and blades, ground to your drawing.
- GrindingAbrasives, superabrasives and grinding wheels, priced and serviced.
Applied Intelligence
An AI grounded in the eight domains of application engineering
Select a pack to see the real attributes it reads and a messy phrase it understands. The packs supply the vocabulary; your own data supplies the truth.
Grounded, not trained
Competence does not come from a bigger model guessing. It comes from your own catalogue, configurators and services, read through the domain knowledge packs.
When Kabaido recommends a part or prices a special, the answer is drawn from the products you sell, the configurators you have set up and the services you offer. The packs supply the vocabulary; your data supplies the truth. Nothing is invented to fill a gap.
Your data stays yours. There is no training on customer data, and our AI provider is configured for zero retention.
Engineered for cost as well as accuracy
Retrieval is database-native and AI is applied surgically.
Kabaido runs a two-model design. Reading and structuring runs on claude-haiku-4-5 by default, and the pipeline escalates to claude-sonnet-4-6 only when a request needs deeper reasoning.
This is why credits are priced the way they are. See pricing.
Where the answers come from, and where they go
The reading is one step. Catalog supplies the matches, Configure prices the specials and Servicing prices the jobs.
Catalog
Where the matches come from: schema-based product data and faceted search, from fifty products to a million.
Configure
How a matched line is priced: configurators with real parameters, rules and prices that explain themselves.
Servicing
How a service job is priced: calculators, inspection gates and serialised items for regrinds, recoats and calibration.