Before we build anything, we spend one to two days on site. The purpose is not to sell a roadmap. It is to find the two or three places where a working system would move a P&L line the executive team already watches, and to size them in the company's own numbers. Nothing leaves the building, and the output is a short document the CFO can argue with.
What we look at
We follow the work, not the org chart. A scan usually covers four flows: how a quote or an order comes in and gets priced, how a customer or a claim gets served, how spend gets committed and paid, and how the month gets closed. In each one we ask the same questions. Where does a person re-key something a system already knows? Where does a decision wait for the one expert who understands the rule? Where does money leak because a check is too expensive to run on every transaction?
The patterns that show up almost every time
The expert bottleneck. A quoting desk, a technical inquiry queue or a claims team where throughput is capped by two or three people who know the catalogue or the policy. Everyone else waits. The payback here is capacity: more orders, faster answers, no new headcount.
The unmanaged tail. Procurement manages the top fifty suppliers well and the next four hundred not at all, because nobody has time to negotiate a contract worth thirty thousand a year. Added up, the tail is often the largest unmanaged spend in the company.
The check nobody runs. Invoices are matched to purchase orders, but not to contract terms, rebates or service levels, because that check takes a person twenty minutes and there are ten thousand invoices. Leakage of a few percent on tens of millions of spend is common, and it is entirely recoverable.
The reconciliation month. Finance spends the first ten days of every month explaining variances by hand across entities and cost centres. The information is in the ledger. The labour is in the narrative.
What the scan produces
A ranked list of use cases, each with a baseline measured from the company's own data, a target the executive sponsor is willing to be held to, the systems involved, and a candid view of the risks. We also say which one we would start with and why, and roughly what a pilot would cost. If none of it clears the bar, the scan says that too.
Why so short
Because the alternative is a twelve-week assessment that produces a deck. The scan is short because the questions are specific and the data already exists. Deciding what to build should take days. Building it well, and proving it in production, is where the time belongs.