AI Tool Selection: Start With What's Already on Your Desk

You're in a leadership meeting. Marketing mentions a new AI writing tool they saw on LinkedIn. Ops says their team started using something they found on their own. Finance wants to know why the company is paying for Copilot licenses that "nobody uses." And now the CEO is asking if maybe you should be looking at that platform a competitor just adopted.

You're not behind. You're just standing in front of a very full kitchen gadget drawer.

Every household has one. The avocado slicer. The garlic rocker. The spiralizer that got used only once three years ago. Each one looked incredible in the ad. And each one is now buried under the next impulse buy while your chef's knife (the one that can do 90% of what those gadgets do) sits right there in the block, just waiting.

That drawer is what AI tool selection looks like inside most organizations right now.

Your Chef's Knife Already Lives in Your Tech Stack

We get asked almost every single meeting "which AI tool is best?".  We also have the same answer: the best AI tool for your organization is almost always the one that's already integrated with your data, your security, and your workflows.

If your company runs on Microsoft 365, that's Copilot. If your company runs on Google Workspace, that's Gemini.

These aren't just AI tools. They're AI tools that already know where your files live, who's in your org chart, and what your team talked about in last Tuesday's meeting. More importantly, they are tools you are likely already paying for and have already been approved from a security perspective. 

That matters more than any feature comparison chart. A standalone AI tool might score higher on a benchmark, but it can't automatically pull from your file repository, summarize your calls, or draft an email reply using context from a thread it's already part of. The tool inside your workflow will always outperform the one that requires your people to copy, paste, and context-switch into a separate app. 

Before anyone opens a new tab to evaluate the next AI product, start here: what can your existing tool already do that you haven't tried yet? For most organizations, the answer is a lot.

When New Needs Show Up (and They Will), Interview Them First

AI tool selection gets tricky when teams start surfacing new needs. A manager wants help building dashboards. A recruiter wants AI screening. A project lead wants automated status updates. The instinct is to go find a tool that does that specific thing. There are a lot of fancy sales people out there promising AI miracles.  

 But, stop right there.  

Before you start shopping, get specific about what the need actually is. I've seen teams ask for a "better AI tool" when what they really needed was a better prompt. Or a workflow adjustment. Or five minutes of training on a feature that already existed in the platform they were paying for.

Here's a prompt you can use to run that conversation with your team. Paste this into Copilot or Gemini and use it as an interview guide with any department that says they need something new:

Try This Prompt

Act as an AI capabilities analyst. I'm going to describe a workflow challenge my team is facing. For each challenge I share, do three things: (1) Identify the specific outcome we're trying to achieve, not the tool we think we need. (2) Suggest how we could solve this using [Microsoft 365 Copilot / Google Gemini] capabilities that already exist today. (3) Only if the existing platform genuinely cannot solve it, flag it as a gap and explain what capability is missing. Start by asking me to describe the first workflow challenge.

This prompt does something important: it forces the conversation away from "what tool should we buy" and toward "what are we actually trying to solve." Nine times out of ten, the answer is already sitting in your existing platform. You just need to find it.

Close the Drawer, Open the Playbook

Every new AI tool your organization adds costs more than the license fee.

There's the security review, the IT onboarding, the data governance questions, the training, and (the one nobody budgets for) the cognitive load on employees who now have to figure out which tool to open first. I've watched teams with access to four AI tools end up using none of them well. That's not an adoption failure. That's AI tool selection working against you instead of for you.

Here's a playbook that actually works:

• Maximize what you have. Give your primary platform a real chance. That means training, use cases, and at least 90 days of intentional adoption before evaluating gaps.
• Document the real gaps. Be specific: "We need AI-generated dashboards from live SQL databases" is useful. "We need better AI" is not.
• Check the roadmap. What doesn't exist today might land next quarter. A quick check of Microsoft's or Google's release notes could save you a six-figure procurement cycle.
• Scope your exploration team. A small, cross-functional group with permission to test new tools is healthy. Think of them as the one person in the house allowed to browse the kitchen store. They might come home with something useful. But they're not authorized to knock out a wall and remodel the kitchen based on what they saw on aisle seven.

If your team is struggling to figure out where to start, use the BORE framework: Boring, Overhead, Routine, or Easy to delegate. Those four categories are the fastest way to find AI-ready work inside any role. Nobody misses formatting a weekly status report by hand.

The organizations I've worked with that get AI adoption right aren't the ones with the most tools. They're the ones that went deep with one tool before going wide. They learned their chef's knife before they opened the gadget drawer.

Your team doesn't need twelve tools. They need one tool they actually know how to use, a clear process for identifying gaps, and the discipline to not chase every shiny thing that shows up in their LinkedIn feed.

What would happen if your team committed to mastering what's already on the desk for the next 90 days?