Intake Agent

Turns a purchase request into a governed PO path with budget, catalog, and approval rules applied up front.

  • Required fields, cost centre, and commodity code before routing
  • Catalog or contracted pricing vs free-text line descriptions

Watch it catch a real case.

Synthetic documents, the same steps the product runs. Pause or replay any time.

ProcureAI · Intake Agent

Split purchase request under approval limits

The document arrives. Nothing has been read yet.

Purchase requisitionPR-7719 bundleFacilities · Illustration · demo dataWaiting
  1. RequesterL. Martins · Cost centre CC-440
  2. PR-7719-A — Laptops (8) subtotal $7,920
  3. PR-7719-B — Docking stations (8) subtotal $2,160
  4. Combined need date28 September 2026 · Same ship-to
  5. Buyer approval limit$10,000 per PR
  6. Note“Please process as two PRs for speed.”
Outcome

Return to requester for consolidation

A person makes the final call

What it checks.

Purchase request intake automation.

  • Mandatory dimensions: requester, ship-to, cost centre, and GL coding
  • Catalog match or three-quote justification when off-catalog spend exceeds threshold
  • Approval matrix based on amount, category, and supplier risk tier
  • Duplicate request detection against open PRs and recent POs

Try it on your document.

  1. 1

    Open the ProcureAI trial and select purchase request, or paste a sample requisition.

  2. 2

    The Intake Agent normalises line items and checks them against catalog and policy rules.

  3. 3

    Submit to your approvers — the agent does not issue a PO without their sign-off.

Your team stays in control.

  • Routing suggestions follow your matrix; approvers can override with a logged reason.

  • No PO or ERP write occurs until an authorised approver confirms.

  • Synthetic demo data never touches your production tenant.

Why it matters.

Purchase request intake is where spend leaks: wrong codes, missing approvals, catalog workarounds. Automating intake does not mean bypassing governance. It means applying your rules consistently before a buyer spends time on the request.

The Intake Agent reads the request, normalises lines, and checks them against catalog, budget, and approval data you already maintain. Approvers receive a complete packet instead of an email thread.

Try it on a dummy requisition to see how Stellitron Intelligence quotes policy and catalog evidence while keeping the approval chain in your hands.

Worked example on demo data

A requester asks for 200 ergonomic chairs off-catalog at $480 each for a branch refresh. The Intake Agent flags the amount over the $25,000 desk-and-chairs policy threshold and asks for a competitive quote or a catalog substitute.

It attaches the policy paragraph and shows comparable catalog SKUs at $410 each, with the price delta spelled out.

The department head still chooses whether to proceed; the agent only blocks routing until the justification field is complete.

Questions buyers ask.

Does intake connect to our ERP?

Pilots integrate with the systems you specify. The public trial simulates routing on Summit Regional Group data.

Can employees use natural language?

They can describe the need in plain text; the agent structures lines for your catalog and coding rules.

How are maverick buys caught?

Off-catalog lines, missing quotes, and duplicate requests are flagged before approvers see the packet.

Is this only for indirect spend?

The same intake pattern works for any requisition your policy defines; scope is agreed during discovery.

A GOOD PLACE TO START

Bring the ambition.
We’ll build the way forward.

Let’s talk about your project

A conversation about your goals, constraints, and next step.