Twenty one tools, and no recommendation

A manual process with compliance attached

As an exporter the company issues certificates of origin across three regions. The steps behind them were manual, with heavy reliance on spreadsheets, which the analysis records as a high risk of error and therefore of compliance failure and penalties.

One specific weakness was named. A bot pulls data from the enterprise system and performs the calculation, but it cannot qualify a preferential status when data is missing, and it requires a person to copy a tariff code from a master spreadsheet and paste it in to find equivalent materials. The risk in that sentence is somebody copying the wrong code.

Scope was set to the two regions carrying the workload. The third was assessed as having no current need, which is a scoping decision worth making explicitly rather than buying a global licence by default.

Twenty one down to five

The market was scouted for twenty one candidate systems and narrowed to five on solution maturity, record keeping, tracking, integration to the enterprise and sourcing platforms, location of operation, and references from the chemicals sector.

The market turned out to be niche in Europe, with few players offering a genuinely global solution because the software is cost intensive to build, and a set of local providers covering Europe and the Americas separately.

Four of the five give global trade agreement coverage. The fifth covered only Europe at the time, with North America on a roadmap, and carried an implementation timeline of up to two years including lead time. That is a roadmap claim, not a delivered capability, and it is the kind of distinction a shortlist has to make out loud.

The functionality is nearly identical, which is the finding

All five support both a web portal and an email route for requesting a supplier declaration, plus a print option for small suppliers. All five do mass request sending, template maintenance, progress tracking with automatic reminders, multi level bill of material calculation, multiple currencies and multiple calculation methods. All five provide full audit trails of the formulas and results.

The analysis says so directly: the overall functionalities are similar. When five products do the same things, a features comparison cannot choose between them, and pretending otherwise produces a confident recommendation resting on nothing.

The one genuine design choice is the portal versus email question, and it cuts against itself. The portal automates properly, because the data lands in the system ready for the origin calculation. Email is what the suppliers actually prefer, and with email the declarations arrive in individual inboxes and get downloaded by hand, which reintroduces the manual work the project exists to remove.

Why no winner was named

Three of the five would not give a price. Two would. Two of the five carry references from comparable industrial companies in the same region. One had a look described as primitive. One cannot currently take declarations by email at all.

None of those differences can be weighed without knowing what the company will pay, which requirements are genuinely non negotiable, and whether it wants one global system or two regional ones. Those are client decisions, and the sprint put eight of them in writing rather than answering them by assumption.

What was handed over instead: five candidates with their pros, cons and prices where given, a scored requirement framework across seven groups, recorded demonstrations, the full long list of twenty one with the reason each was rejected, and vendor contacts.

Then five steps in order. Test which requirements actually matter and whether any are deal breakers. Re-rank on price and usability against current and future needs. Take two or three forward to a demonstration on the company own data. Finalise the budget. Build the business case.

There is no measured result here and there could not be. Nothing was bought, nothing was implemented, and no error was avoided during the sprint. What was produced was a shortlist of five from a market of twenty one, and a decision process with the assumptions made visible.