Skip to content 
Six months from now, you will look back at today. The question is whether you will see real progress or another round of meetings about what AI might do for your organization someday.
Over the past few weeks I have written a lot about personal AI agents and how they are changing the way individuals work. The response has been fantastic. But the conversations that followed kept coming back to a bigger question from leaders: “That is great for me. What do I do with my organization?”
That is what this post is about.
The tension every leader is feeling
If you lead an organization, you are probably feeling pulled in two directions.
On one side, there is pressure to “do something with AI.” Your board asks about it. Your staff are already experimenting with it, with or without a policy. Every vendor email promises transformation.
On the other side, there is reality. You are running on systems that were bought a decade ago, spreadsheets that only one person understands, and integrations held together with good intentions. Budgets are tight, and your team is already stretched serving your mission.
So the choice can feel like this: launch a big, expensive transformation you are not sure you can finish, or do nothing and fall further behind.
I want to offer a third path. Real AI value does not come from a big bang. It comes from steady, incremental improvement on a clean foundation. AI is only as good as the data and processes underneath it. Fix those a little at a time, and every AI tool you adopt gets more useful.
Here is where I would start.
The 5-part checklist
- Pick one painful process and measure it. Do not start with technology. Start with the process that frustrates your staff, your clients or customers. Write down how long it takes today, how many hands it passes through, and where it breaks. That baseline is what lets you prove progress later, to your board and to yourself. This week: Ask three frontline staff which task wastes the most of their time, pick one, and time it end to end.
- Clean up your data and decide where the truth lives. AI tools cannot fix messy data. They amplify it. If donor, client, or constituent records live in five places, you need one source of truth, and for many organizations that is a CRM such as Salesforce. Name an owner for your core data, agree on what “clean” means, and start with the records that matter most to your mission. This week: Pick one key record type, such as contacts or cases, and count your duplicates and blank critical fields. That number is your starting point.
- Audit your legacy tech: keep, connect, or retire. You do not have to replace everything, and you should not try. List your major systems and sort each one into three buckets: keep it as is, connect it to your source of truth, or plan to retire it. Most of the value is usually in the middle bucket, getting systems to share data instead of relying on manual re-entry. This week: Build a simple one-page inventory of your top ten systems with the owner, the cost, and a keep, connect, or retire label for each.
- Run small 30-day pilots with a named owner. This is where AI comes in, and it should come in small. Pick one use case, such as an AI assistant that helps staff summarize case notes, or answer common internal questions. Give it one owner, a clear success measure, and 30 days. At the end, decide to scale it, adjust it, or stop it, and treat all three outcomes as a win because you learned something. This week: Choose one pilot, name the person accountable, and write down in a single sentence what success looks like on day 30.
- Set a monthly improvement rhythm and track your wins. Incremental change only works if it is continual. Set a standing monthly meeting, 45 minutes is enough, to review your baseline numbers, pilot results, and the next small improvement. Keep a running list of wins, even small ones. That list becomes your story for the board, your funders, and your staff. This week: Put a recurring monthly “improvement review” on the calendar and start a shared wins log with the first entry from this checklist.
What 6 months from now looks like
If you start this week and keep the rhythm going, here is what you can realistically expect to see when you look back.
- One process measurably better. You will have before and after numbers for the process you picked, such as hours saved per month or reduction in turnaround time.
- Data you can trust. Your core records will be cleaner, owned, and living in one place, which makes reporting faster and AI tools far more reliable.
- A clear technology roadmap. Instead of a vague worry about legacy systems, you will have a documented plan for what stays, what connects, and what goes.
- Several pilots behind you. You will have run several 30-day pilots, with real evidence about which AI tools help your staff and which do not.
- A culture of steady improvement. Most importantly, your team will be in the habit of making things a little better every month. That habit outlasts any single tool or vendor.
None of this requires a massive budget or a multi-year program. It requires focus, ownership, and consistency.
Start small, start now
The organizations that get the most out of AI will not be the ones that bought the biggest platform. They will be the ones that cleaned up their foundation and kept improving, month after month.
You do not need to have it all figured out. You just need to start.
If you would like a thinking partner as you work through this checklist, whether that is getting your Salesforce data in shape, sorting out legacy systems, or designing your first AI pilot, I would be glad to talk. Reach out to me or the Belmar team at belmarcloud.com.