What does Dock add to SAP Ariba for enterprise sourcing?
Dock is where a sourcing agent records its interpretation of an Ariba RFx event. Ariba still holds the event, the supplier responses, and the award. Dock holds the agent's scorecard, its recommendation, the category manager's decision, and the Workday cost-center context it needed to weight the spend. The pairing lets one agent read a 40-supplier response in a morning and surface a defensible shortlist by lunch, with every line of evidence pointing back to a fresh Ariba read.
The architecture
SAP Ariba and Workday stay the system of record for the raw data: the sourcing event, supplier profiles, response documents, award decisions, and the cost-center hierarchy that owns the spend. Dock is the system of record for what the agent interprets from that data. Each Dock row carries a pointer back to the platform record (ariba_event_id, workday_cost_center_id), agent identity, decision, reviewer, and timestamp. The agent re-fetches Ariba and Workday data via fresh API reads when it needs current state, so a scorecard never drifts from the live event.
The Dock surface: RFP Analysis table
| ariba_event_id | supplier | agent | scorecard_total | recommendation | reviewer | status | reviewed_at |
|---|---|---|---|---|---|---|---|
| EVT-2026-0412 | Northwind Logistics | sourcing-agent-01 | 84 / 100 | Shortlist | g.iyer | approved | 2026-05-29T14:02Z |
| EVT-2026-0412 | Brentford Freight | sourcing-agent-01 | 71 / 100 | Decline (capacity gap) | g.iyer | approved | 2026-05-29T14:04Z |
| EVT-2026-0412 | Halcyon Transit | sourcing-agent-01 | 79 / 100 | Shortlist, flag for compliance review | k.okafor | pending | 2026-05-29T14:06Z |
Every row points back to one Ariba event and one supplier response. The scorecard total is the agent's interpretation, never an Ariba field.
The worked workflow
A category manager opens an Ariba sourcing event for inbound freight and invites twelve carriers. The agent watches for the "responses received" webhook. Once Ariba marks the event closed, the agent:
- Re-fetches each supplier response from Ariba via API.
- Pulls the requesting cost center from Workday to weight the scorecard against last year's actual spend.
- Scores each response against the rubric in /blog/dock-for-procurement (price, capacity, insurance, on-time benchmark, ESG attestation).
- Writes one row per supplier into the RFP Analysis table with
ariba_event_id, the score, and a recommendation. - Posts a single summary message tagging the category manager.
The category manager reviews the rows, approves three for shortlist, and rejects one. The agent then opens the consent gate described in /blog/dock-for-legal: writing the award back to Ariba is a dangerous operation, so it requires an explicit human approval before the API call fires. Until that gate clears, Ariba sees no change.
Why it matters
Enterprise sourcing teams already trust Ariba with the event of record. SAP reports that its B2B network supported over 5.3 million companies and $3.75 trillion in annual transactions as of June 2022 (Wikipedia, SAP Ariba). That scale is exactly why the interpretation layer needs to live somewhere else. You do not want an agent overwriting Ariba fields with its own opinion of "best supplier."
ISM defines supply management as "the identification, analysis, determination, procurement and fulfillment of the goods and services an organization needs to meet short- and long-term objectives" (ISM). Agents are good at the analysis step and bad at the determination step. Dock draws the line cleanly: the agent analyzes in Dock, the human determines in Dock, and only the determination flows back to Ariba.
The audit story is the second half. Every scorecard row carries the agent identity, the reviewer, and the timestamp, which is the substrate the patterns in /blog/agent-audit-and-compliance and /blog/agent-identity-lifecycle depend on. Finance can reconcile the award trail without leaving the same workspace where AP processes the resulting invoices, as covered in /blog/dock-for-accounting.
Try it
Open a Dock workspace, point a sourcing agent at one closed Ariba event, and review the scorecard it writes.
FAQ
Does Dock replace SAP Ariba?
No. Ariba stays the system of record for the sourcing event, supplier responses, and the award. Dock is the system of record for the agent's analysis and the human decision on top of that analysis.
Can the agent award a contract in Ariba on its own?
No. Writing an award back to Ariba is a dangerous operation. Dock requires an explicit human approval at a consent gate before the API call fires, following the pattern in the legal cluster.
How does the agent stay current if a supplier updates a response after scoring?
The agent re-fetches the response from Ariba on every scoring run. The Dock row keeps the ariba_event_id pointer, so a re-score writes a new row rather than mutating the prior one, preserving the audit trail.
Where does Workday come in?
Workday holds the cost-center hierarchy that owns the spend. The agent reads the requesting cost center to weight the scorecard against prior-year actual spend, so a small department does not get a scorecard tuned for an enterprise contract.
