A program coordinator finishes a site review and types her notes into the case management system. Then she enters the same scores into a spreadsheet for reporting. Then she emails a summary to the finance team, and someone there keys part of it into a third system. Nobody designed the process this way. It grew one workaround at a time.
Now picture adding an AI tool to that process. It might write her summary faster. But the three systems still don't talk to each other, and the re-keying is still there.
Many organizations are building next year's budget right now, and AI is often on the list. That's a good instinct. In our work with education, healthcare, and government teams, though, we've seen the same pattern again and again. The biggest gains rarely come from the tool itself. They come from fixing the work underneath it.
So before choosing a tool, it helps to ask two questions. Where does your team lose the most time to manual work? And if you redesigned that work from scratch, where would AI actually have a job to do?
Why adding a tool isn't enough
Manual work tends to pile up in a few familiar places:
- Duplicate data entry. Staff enter the same student, patient, or applicant record into more than one system.
- Manual hand-offs. Information moves between departments by email, shared drives, or a printed form that someone scans.
- Searching for information. People spend hours tracking down the latest version of a document or the right person to ask.
Put an AI tool on top of these problems and the friction usually stays. You end up paying for new technology while staff keep doing the same workarounds. A better approach is to rethink how the work should flow first, then bring in AI where it has a clear job.
A practical five-step approach
This works for an accreditation office, a health agency, or a state program.
- Find where time is being lost. Sit with the people doing the work and ask them three questions. What do you enter more than once? What are you always waiting on? What part of the week do you dread? The answers are usually specific, and they point straight to the problem. While you're there, write down how long the process takes today, how often errors happen, and how many people touch it. Without a starting point, you won't be able to show what improved.
- Rethink the work, not just the steps. Don't only ask "how do we automate this?" Ask "do we need this at all?" Sometimes a report exists only because two systems can't share data.
- Connect your systems so information is entered once. In our coordinator's case, that means her site review scores flow directly into the report and the finance system. No one types them again.
- Add AI where it makes a measurable difference. Once the workflow is clean, AI has a clear job. It can draft the summary, route a request to the right person, or flag a form that's missing information.
- Judge the results against where you started. Compare time, cost, and accuracy to the numbers you wrote down in Step 1. Then check one more thing: do staff actually prefer the new way? If people quietly go back to the old spreadsheet, the project didn't succeed.
Start small this month
AI is a tool, not the goal. The goal is work that runs better. The organizations that get real value from AI usually did the less exciting work first. They mapped the process, connected their systems, and cut the steps nobody needed.
You don't have to tackle everything at once. Before your budget is final, pick one workflow that frustrates your team. Spend an hour on Step 1: talk to the people who do it and write down today's numbers. That alone will tell you whether you need a new tool, a better connection between the systems you already have, or both.
If you'd like a second set of eyes on what you find, we're always happy to talk it through. A small, well-designed win is a strong foundation for whatever comes next.
