Copilot Not Working Well For You? Let’s Explore That.
How To Give Copilot or Gemini Access to Your Business Data
We’ve made real progress in getting Microsoft Copilot to be useful in our office, but it did not happen automatically. One of the biggest early lessons was that Copilot is not magic. If your information is messy, fragmented, or inaccessible, the output will reflect that. As we got our internal content better organized and removed some of the barriers that kept Copilot from seeing the information it needed, the tool became far more useful.
That’s been helpful but, like most businesses, we have important data sitting in systems outside of our Microsoft 365 (or Google Workspace) environment. The information in our ticketing system, phone system, CRM platform, and accounting software is not natively visible to Copilot, so we’ve started thinking about how to expose that data to AI without overengineering the solution on day one.
The good news is you don’t have to start with a complex integration. There’s a much simpler path we’ve been working through as we figure out how to connect the dots.
Step Zero: Build the Right Foundation First
Before we talk about connecting anything external, we need to address the most common issue we see, which is that AI cannot make sense of what isn’t visible and organized. That means we need to identify:
- Information that only exists in someone’s head is invisible
- Information locked in systems outside your AI’s reach is inaccessible
- Information scattered across multiple tools with conflicting versions is unreliable
In our own environment, the first real progress came when we focused on getting our internally generated content into the right places:
- Emails properly categorized
- Documents stored in shared, accessible locations
- Meeting notes and outputs consistently captured
- Permissions aligned so the right people—and tools—could see the same information
Only after we cleaned that up did Copilot start producing results that felt actionable instead of generic.
If you skip this step, none of what follows will work the way you want it to.
Level 1: Email as a Bridge (Fast and Effective)
The goal here is to find the fastest, cheapest, easiest way to connect two systems, and email is often the simplest way to import external system data to Copilot or Gemini. We’ve been experimenting with having summary information from our ticketing and phone systems emailed into a shared inbox where Copilot has visibility.
It’s not a perfect solution, but it works surprisingly well as a starting point in our case. No APIs. No custom development. No complex setup. Just a simple way to get the information to a location where AI can see it.
Level 2: Built-In and Third-Party Connectors
Setting up and forwarding email from outside sources can only take you so far, and in some cases it’s not really practical, so at some point you may need to look into built-in connectors and third-party tools like Zapier. If one of these connectors is available, it can be a simple way to automatically move data between systems or trigger workflows based on events (ticket closed, form submitted, etc.).
Automated connectors are great in that they are usually easy to set up and can be an affordable way to get traffic flowing between two disconnected systems. You do want to pay attention to the way the fees are set up though. Some of these tools charge on a per transaction basis, so if you have a lot of information flowing the fees can add up quickly.
Level 3: Webhooks and API-Based Integration
At the top of the ladder are Developer-heavy tools like Webhooks and APIs. These are where things become more powerful, but also more difficult to set up and maintain. Webhooks and APIs allow developers to create sophisticated integrations between systems that are natively outside of Copilot or Gemini, and can help our new AI friends understand how the information connects.
Developers love these tools (and so do I), but as small businesses, it’s important that we think through the value proposition before setting up an API, since they typically require significant effort to design, install, and maintain.
Progress Over Perfection
When it comes to AI, we are not trying to build a perfect integration today, and it doesn’t really make sense to even try. AI is evolving too quickly to justify overbuilding every integration today. Focus first on making your information organized and accessible, then connect systems in the simplest way that delivers useful results.
This approach makes sense because it addresses a core issue for most small businesses: getting your information organized and accessible. It’s what businesses have needed to do since forever when they are preparing to scale, and it just so happens to be exactly what AI needs to be effective. Connecting our information silos is part of the equation, but spending lots of time and/or money on making it perfect is not worth the effort.
The goal is not to build the perfect integration right away. The goal is to make useful progress with the tools available today.
Final Thought: Don’t Quit Too Early
Developing and integrating Copilot or Gemini into a small business takes time and energy, and at the speed everything is moving it’s a given that you will make some mistakes.
We certainly have.
If you stay focused on organizing your information and improving your processes, I’m certain that the effort will pay off, whether your next step is Copilot, Gemini, or something else.
Let’s Compare Notes
If you’re working on or thinking about what Copilot or Gemini could do in your business this year, I’d love to chat with you about your efforts, and the easiest way for us to connect is for you to book a slot on my calendar so we can connect and talk about how AI is impacting our daily lives.
We’re learning this in real time, just like you are, and I suspect there’s more we can figure out together.