Most integration plans after an acquisition start with the general ledger and the reporting lines. The operational systems, the ones that record work, set rates and send invoices, tend to be left running side by side. That is reasonable for a while, until nobody can say what is running in parallel or who keeps it in step.
None of the checks below needs a project. Each is a question someone can answer this week, from the systems as they are.
The seven gaps
1. Separate billing for each brand
Each acquired business still raises its own invoices, from its own system, on its own templates and payment terms. Clients who buy from two brands receive two invoices, chased separately.
Ask this week: how many systems raise an invoice today, and does anyone hold a single view of what one client owes across all of them?
2. Client and worker records duplicated
The same client exists in several systems with slightly different names and terms. The same contractor may appear twice, with two rates and two sets of bank details.
Ask this week: if you search for your ten largest clients in each system, how many appear more than once, and which record does finance treat as correct?
3. Rate tables and project codes that differ
One business bills by role, another by grade, a third by project. Project codes differ too, so nothing rolls up without a mapping in somebody’s spreadsheet.
Ask this week: is there a written mapping between the rate cards and project codes of each business, and who updates it when a rate changes?
4. Approvals living in email
Discounts and rate exceptions are agreed in email or chat, then keyed into the system by someone else. The system shows the result but not the decision, so an audit question turns into an inbox search.
Ask this week: for the last three non-standard deals, can you find the approval in a system, or only in somebody’s inbox?
5. Two sources of truth for revenue
The operational system gives one number for the month and finance gives another. The difference is explained every month by the same person with the same spreadsheet.
Ask this week: when the two revenue numbers differ, is the reason written down, and is it the same reason every month?
6. Integrations held together by one person
Data moves between the acquired systems through scheduled exports, scripts or a connector that one person set up and one person understands. When that person is on leave, the data stops or quietly goes wrong.
Ask this week: if the person who maintains each integration were away for two weeks, who would notice a failure, and who could fix it?
7. AI pilots started on unclean data
A pilot has started to summarize contracts, forecast demand or match timesheets, using data from systems that still disagree with each other. It produces confident answers built on duplicated records and mismatched codes.
Ask this week: which data does each AI pilot use, and has anyone checked it against the gaps above?
What to do with the answers
A few honest answers usually show where the money is. Not every gap needs closing in the first hundred days. The point is to know which ones are costing you, so that leaving a gap open is a decision rather than an accident.
Records and rates come first, because billing, reporting and automation all depend on them. AI belongs at the end, once the process underneath is clean.
If you would like the costed version, the Throughset Diagnostic takes one process across one or two systems and, in five working days, produces a line-item list of what is broken between them, what it costs to leave and what it takes to fix. It is a fixed fee from USD 3,500, payable on completion, and you keep the findings either way. It starts with a free twenty-minute scoping call.