Projects go sideways when everyone assumes the same sentence means the same thing. “We need a portal” can mean a brochure site, a client account, an internal operations tool, or a full replacement of how work is done. Preparation is how you find out which one you mean.
Bring the current process, even if it is imperfect. Screenshots of the spreadsheet, a sample of the email thread, the form staff actually use — these are more valuable than a vision deck. They show the exceptions, which is where software usually fails.
Name the users with some precision. “Admin” is not a user. The person who assigns jobs, the person who approves spend, the client who checks status, and the manager who needs a weekly view often need different things. If they are collapsed into one role, the first release will frustrate at least two of them.
List the systems that already hold truth: accounting, CRM, payments, identity, file storage. Integration is not a later nicety. If the new software cannot talk to these, people will copy data by hand and quietly abandon the new tool.
Be explicit about constraints. Budget range, decision dates, compliance needs, who must be consulted, and whether there is an internal technical owner. Constraints do not scare a serious partner. Hidden constraints do.
Finally, define success in operational language. “Modern” is not a result. Fewer lost requests, faster month-end, clients who can self-serve a status, or a backup you have restored — those are results. Pentacom uses discovery to turn that preparation into a scope you can approve with your eyes open.