Another AI announcement. Another tool. Another video promising to change how you work.
You save it for later. Then you save three more.
Before the workday starts, you already feel behind.
For independent insurance agencies, the challenge is choosing where AI can make a difference and following through.
Start with one question: What problem is worth solving this week?
Your agency might have several AI subscriptions, a few unfinished pilots, and employees experimenting every day.
What has improved?
If that’s hard to answer, it’s time to focus. Before trying another tool, ask:
Choose one task. Keep it manageable, and make sure solving it would matter.
We were using AI regularly at Catalyit, testing tools and exploring ideas. But partner reporting remained a recurring headache.
We report to more than 100 partners. Each report required gathering information from our CRM, Google Analytics, and other sources. Our systems were connected, but the work still involved plenty of manual effort.
We started with HubSpot’s AI, then connected HubSpot to Claude.
Several team members tested different prompts and compared results. We looked at what each approach captured, what it missed, and which instructions produced the most useful information.
Then we combined what worked into a reusable skill.
We could generate a detailed report consistently. But we had another problem: it was 10 pages long.
Who was going to read all 10 pages?
We needed a report partners could scan and understand quickly. Using our design system in Claude, we brought the key findings forward, applied our branding, and cut the report to a couple of pages.
That gave us a repeatable process and a more useful result.
The win also helped the team see another opportunity. We began using Claude and the same design system to build email communications for distribution through HubSpot.
One solved problem gave us the confidence to tackle the next.
Your agency doesn’t need to do everything at once.
If you’re watching: Inventory the AI tools your team already uses and establish a written AI policy.
If you’re experimenting: Pick one workflow. Record how long it takes and what a good result looks like. Test AI several times and compare.
If you’re using AI regularly: Assign an owner, document the process, and decide who reviews the output.
If you’re measuring results: Use the evidence to decide whether to adjust, expand, or stop.
If you’re scaling: Build on what has demonstrated value. Keep checking quality as more people use the process.
Each stage should lead to a clear next action.
A vendor or consultant should ask about your agency before recommending a solution.
What slows your team down? Which systems do you use? What have you already tried? What result would make the investment worthwhile?
Expect those questions. If the conversation jumps straight into a pitch, bring it back to the work you need to improve.
Check your existing tools, too. You may already be paying for the capability you need.
Give your team 30 minutes to discuss what they’ve tried, what worked, and where they’re stuck.
Then complete this sentence:
“This week, I’ll try AI to help me with ______.”
Choose an owner. Define success. Test the process, review the result, and adjust.
The next AI announcement can wait. Pick one problem that matters to your team and work toward a result you can measure.
Take a breath. Find your focus. Make your next move.
You don’t need to have AI figured out. You need one useful next move. Catalyit created a free worksheet to do just that.
Use it to pick one task, try it, and see whether it helped.