A NocTeam MVP in a one-company internal pilot, informed by 20 customer interviews and five prototype tests.
Twenty customer interviews and five prototype tests informed the shift from dashboards to reconciliation and evidence-linked alerts.
Definition & measurement context ↓Small-carrier operators and back-office teams reconciling invoices, payments, factoring and bank records.
The original direction assumed operators needed a better dashboard. Before they could use one, they had to bring fragmented financial and operational records into agreement.
This is NocTeam work, in a one-company internal pilot. Fifteen-plus months describes historical data, not live pilot duration. The 20 discovery interviews are not a pilot-user count.
Across 20 customer interviews and five prototype tests, I found that operators first struggled to bring together data from email, QuickBooks, screenshots and spreadsheets. A dashboard would still leave that setup work with the customer. I redirected the MVP toward reconciliation and evidence-linked alerts with engineering and design.
We defined the order-to-cash workflow across invoices, bank activity and factoring records, using OCR, deterministic rules and an LLM fallback. I narrowed the MVP around a detected issue and its supporting records so an operator could decide what to do next.
Focus on evidence-linked operational alerts. Research shifted the problem definition from displaying information to making fragmented information actionable.
I replaced dashboard-first direction with reconciliation and evidence-linked alerts.
A payment memo references the invoice, but the ledger has not been reconciled.
Suggested next stepInspect the records before changing invoice status. The operator owns the decision.
Awaiting operator review
Separate demo states illustrate why action and confirmation need separate event counts.
Project-reported outcomes are separated from discovery counts and program scale. Definitions and measurement limits are included below.
Twenty customer interviews and five prototype tests informed the shift from dashboards to reconciliation and evidence-linked alerts.
Invoice-to-payment matching result reported from the one-company pilot. The correct-match criteria, reference labels and evaluated pair count are not available, so the percentage is kept in measurement context.
Correct-match definition, sample size, ground-truth process and live pilot dates unavailable. Not general detection accuracy.
Method: Project-reported matching evaluation. Population: One carrier company; 15+ months of historical operating data. Period: Not documented.
Share of alerts prompting action or confirmation, combined.
Action and confirmation are not separately counted. Not a share of users or a weekly active-user metric.
Method: Career-master alert-level outcome. Population: One-company internal pilot. Period: Not documented.
Reported actionable issues surfaced weekly in the pilot.
Live duration, issue adjudication and week count unavailable. Historical-data coverage is not live adoption duration.
Method: Project-reported usefulness outcome. Population: One-company internal pilot. Period: Not documented.
Expansion is a next validation step, not an achieved rollout. A second carrier and separate action/confirmation events would test whether the product thesis survives beyond the initial company.