Bring the BPO work home with digital workers.

Requisition management comes back in house, run by digital workers you control. Every request reviewed, the buys in scope negotiated, no added headcount.

Get a demoHow every request is reviewed
  • 1,000,000+requisition checks since March 2026
  • 60+global enterprises
  • <1 dayfrom request to purchase order
  • 90%cut in requisition cycle time at SUEZ

Every request reviewed in house

The Alignment Agent reviews every request against your rules the minute it is raised, at any volume.

  • Every request reviewed the moment it is raised, with no sampling
  • A request gets the same review whether it is raised in chat, a portal or your P2P
  • The requester gets the fix in the same conversation, so a clean requisition goes through
Alignment Agent
<1 minto check a request against every policy you have
Illustration: a policy review of one requisition. The Alignment Agent corrected a quantity itself, sent the payment terms back to the requester because the quote and the contract disagreed, passed the supplier record and the coding, and the requester corrected and resubmitted nine minutes later

Negotiation on the buys nobody had time for

The Tactical Sourcing Agent runs the quotes and the negotiation on the buys that used to go straight to purchase order, whatever their value, and your analysts make the award.

  • Quotes collected and compared on every request in scope
  • Competitions run on buys that used to go straight to purchase order
  • The award stays with your analyst
Tactical Sourcing Agent
11%average saving on tactical buys that closed with a saving
Illustration: the Tactical Sourcing Agent running a queue of RFQs at once, with counts of requests in market, awarded this week and waiting on a buyer

Prove it beside the BPO first

Digital workers review the same requests as your provider in observer mode, so you see how often they agree and decide the renewal on the numbers.

  • Same requests as your provider, with no effect on any of them
  • Agreement, false alarms and misses shown before it acts
  • Every exception goes to the owner you name
Cost it against your BPO
99%agreement with a customer's own reviewers before it acted alone
Illustration: a proof panel for one rule, payment terms match the contract, version 4, currently advising. The observer run reviewed 1,240 live requests with no effect, the backtest ran against 3,100 requisitions the customer's own reviewers had judged, agreement with those reviewers was 97.5% with 1.6% false alarms and 0.9% misses, and the rule is ready to switch on

Digital workers for the shared services center

Six digital workers that do the outsourced work. The center decides which run and where.

  • Intake Agent

    Turns a request in plain words, in Teams or your P2P, into a clean, correctly coded requisition and hands it on.

  • Alignment Agent

    Turns your policies and strategy into one rulebook, checks every request against it and gives every digital worker its rules.

  • Single-Source Negotiation Agent

    Negotiates every single-source requisition before the PO exists, for better price, payment terms and working capital.

  • Tactical Sourcing Agent

    Runs the RFQ, drives the competition and negotiates the final terms on buys your team never had time to run.

  • Advanced Sourcing Agent

    Runs the multi-round events your buyers used to carry alone, built on your category strategy.

  • Post-Sourcing Negotiation Agent

    Takes the bids from your closed RFx and negotiates a final offer with every shortlisted supplier before you award.

“Days of manual reviews became minutes. Faster decisions, better compliance, and real efficiency gains.”
Rob Baillie · Head of Procurement · SUEZ UK

What changed for other centers

Case studies and articles where the people running the volume tell you what happened.

Questions shared services leaders ask about autonomous procurement

How does a shared services center reduce the cost per procurement request?

A shared services center reduces the cost per request by having AI agents check every requisition the minute it is raised, so volume grows without the analyst count growing with it. Pactum's AI agents have run 1,000,000+ requisition checks since March 2026. The center keeps its own rules and stays the single control point for the company.

Should requisition checking be outsourced to a BPO or automated?

Requisition checking can be automated so every request is checked, where an outsourced team often checks a sample. AI agents check every request against the center's own rules. You can run them in observer mode alongside an existing outsourced team, compare the two on the same requests, and make the contract decision on the numbers.

How fast is an automated requisition check?

An automated requisition check takes minutes rather than days. At SUEZ, requisition cycle time was cut by 90%. At Pactum's deepest deployment, a request reaches a purchase order in under a day.

Who owns exceptions when an AI agent checks requisitions?

The center owns exceptions and decides what happens when a check fails: advise the requester, escalate to a named owner, or reject with the reason attached. A requisition is never left blocked, because anything an AI agent cannot settle goes to a person with the finding and the fix.

Can AI agents catch requests that bypass the central procurement team?

Yes, AI agents check requests that would have gone around the center. A request raised in Teams or straight in Coupa gets the same check, so it is caught at the request and brought back through the center. Every channel gets the same check, so none is left unchecked.

Does Pactum work with the center's existing Coupa or SAP Ariba setup?

Yes, Pactum works inside the setup the center already runs. Teams is the front door, Coupa or SAP Ariba stays the record, and Pactum does the work in between. Write-back and the logged decision trail are documented for Coupa, and SAP Ariba is supported as a named approver step. Pactum is SOC 2 Type 2 compliant and ISO/IEC 27001:2022 certified. See integrations.

What happens to the center's analysts?

The routine checking goes and the judgment work stays with the team. Analysts move to the exceptions, to the decision points on sourcing events, and to operating the AI agents for the whole company. At SUEZ, 80% of manual reviews were eliminated.

How do we get started with Pactum?

You get started with Pactum on one category, one region or one business unit, in three steps: observer mode, advising and acting. In observer mode, AI agents check every requisition in Coupa or SAP Ariba and change nothing, so you see where requests break policy. When advising, they tell the requester what is wrong and how to fix it. When acting, they approve, fix or escalate requests once your team agrees with their verdicts.

Which spend does Pactum work on?

Pactum works on indirect spend: the goods and services a company buys to run itself, such as IT, facilities, MRO and professional services. AI agents run each request end to end, from intake and the policy check to sourcing, negotiation and the purchase order, inside Coupa or SAP Ariba where each agent is live. That includes the requests outside your catalog and contracts, where no price was agreed in advance.

See a day of BPO work done by digital workers.

Bring the requisitions your provider handles and the rules they apply, and we'll run the demo on them.

Get a demo