Aftawork AI Labs

About IndigiMatch

A controlled research and matching platform for Indigenous vendor capability. Neither a directory nor a search engine: a governed system in which every conclusion about a vendor is produced by published rules, recorded with its evidence and date, and reproducible later by someone who was not in the room.

The Wayfinder story

Long before satellite positioning, the nations of the Great Lakes navigated by the stars. A sky is not a map; it is thousands of points, most of them irrelevant to the journey in front of you. Wayfinding is the skill of reading them together, recognizing the pattern that matters, and following one path to a fixed destination. It is knowledge held in the community, tested every season, and passed on with its reasons.

That is the work IndigiMatch does. The sky is the evidence base: the federal Indigenous Business Directory, the national certification registry, contract award history, controlled goods registrants, training and workforce data, and years of field research on live contracts. The path is the research method, lighting up one verified step at a time. The destination is a vendor who can do the work, with the proof beside the name, and the reflection on the water is the record left behind, so the route can be followed again.

The identity draws exactly that: a night sky over Lake Superior at Fort William First Nation, where the platform was built. Every colour in it is the sky and its reflection in the water from dusk to dawn, and every verification tier is a state of the same mark, from a faint point to the gold star. It is a way of saying, without a word, how much evidence stands behind a recommendation.

What it is

The question IndigiMatch answers is narrow: for a defined scope of work, which Indigenous businesses can genuinely do it, on what evidence, and where none yet can, stated honestly. It is a proven system in an operational environment, in use on two Canadian aerospace in-service support programs and demonstrated against a live $550 million federal defence infrastructure procurement.

Three habits shape everything it produces

  • Verification depth is visibleVendors are ranked by how thoroughly they have actually been verified, from full depth on a live contract down to a directory entry. A reader sees the strength behind a conclusion, not just a name.
  • Capability is assessed per requirementA vendor proven in one capability is not treated as proven in another. Fit is scored fresh against the scope in question, every time.
  • Gaps are reported as findingsWhere the Indigenous supply base does not hold a capability, the platform says so and evidences it. That is often the more useful half of the output, because it turns a participation target into a capacity-building plan.

How it works

Registries and records go in; verified, dated, reproducible findings come out. Between them sits a governed workflow with two human approval points that cannot be delegated to software.

How IndigiMatch worksInputs on the left: registries, contract history, client scope, and vendor intake. A governed workflow in the centre: identify, research in progressive depth, score, team, human approval, lock and log. Outputs on the right: verified vendor profiles, reports, teaming pathways, gap register, and audit record.InputsGoverned workflowOutputsIndigenous Business Directoryfederal registry; governs PSIB eligibilityNational certification registrycorroborating ownership evidenceFederal contract award historyand controlled goods registrantsTraining and workforce datasetsboth official languagesClient scope of workdirect and indirect work packagesVendor readiness formssubmitted by Indigenous businessesIdentifycandidates per work packageResearch in progressive depthidentification to full-evidence lockScore0 to 100 on one published rubricTeampairings where no single vendor holds the capabilityLock and logevery record dated with its evidencehuman review gatehuman approval gatePublished, tested decision rules held as data. AI assists research under those rules;it does not set scores and cannot grant approvals.Verified vendor profilesseven sections, tiered and datedSix report typesmarket research to complianceTeaming pathwaysevery viable pairing, one formulaGap registereach with an owner and a pathAudit recordreproducible by any reviewerPersistent knowledge baseclosed contracts inform the next

Research in progressive depth

Each capability is researched in stages, from identification and screening through live research to a lock at full evidence depth, and the depth reached is recorded on the face of the work. No data enters a recommendation until it has been validated more than once. Research runs against the governing federal and national registries, federal contract award history, the controlled goods registrant directory where the work requires it, and training and workforce datasets, in both official languages.

Scored on one rubric, teamed by rule

Every vendor is profiled across seven sections: company overview, technical capability, match assessment, Indigenous alignment, capacity and risk, outreach status, and strategic value. Fit is scored from 0 to 100 on a published rubric, with a preference for federal registry status. Where no single vendor holds a capability, every possible pairing is enumerated and scored on the same formula, and no pairing is written until it passes the compliance gates. Joint ventures are built by rule, not by relationship.

Controls that can be shown, not just described

  • Published, tested decision rulesHeld as data and changed only through a review gate, never silently. When the federal ITB policy was modernized, the platform absorbed the transition through one governed review pass without changing a formula or a line of code.
  • Two human approval pointsNeither can be delegated to software. An approved AI service assists with research and interpretation under the rules; it does not set scores and cannot grant approvals.
  • A dated, reversible change recordEvery research action, teaming pass, and format operation is logged with its timestamp, depth, and rationale, for the life of the contract.
  • Freshness controlsEvery dataset carries its refresh date. Out-of-date information cannot silently feed a live recommendation, and every conclusion is stamped with the rules and data vintage under which it was reached.
  • Honest blanksWhere a check cannot be completed, the platform records that fact and its reason instead of leaving a blank that later reads as a clear result.

What it does not do

  • It does not certify Indigenous ownership. It verifies against the governing registries and records what it finds; certification remains with the certifying bodies, and eligibility determinations remain with the Crown and the program's own criteria.
  • It does not guarantee a vendor's performance. It evidences capability and standing at a point in time.
  • It does not replace relationships. Identification is not engagement, and the platform treats the two as distinct stages.
  • It does not produce ITB or participation credit. It produces the vendor evidence base; what is claimable, and how, is a matter for the Crown and the policy.
  • It does not evaluate bids. Its scores support market research and vendor engagement; award decisions belong to the buyer.
  • It is not yet a hosted service. It runs as a controlled system for a single contract at a time. A multi-client hosted version is in development and has not been released.

Ownership and operation

IndigiMatch is owned by Aftawork AI Labs. It was founded by Timmy Marvin Pelletier, a member of Fort William First Nation, and built at his office on Fort William First Nation, Ontario, from two decades of procurement experience on defence and infrastructure programs. It is delivered on client contracts either directly or under licence to an operator working with the prime contractor.

Data supplied by vendors is used only to verify ownership and record capability. It is not sold, and it is not shared beyond the engagements a vendor is matched to.

welcome@indigimatch.ai