The 6 Organizational AI Milestones Every Company Needs to Hit

From the Automate Your Agency Podcast — EP 107

Everybody's using AI these days. But there's a big difference between individuals experimenting with tools and an organization that has actually built AI fluency into the way it operates.

In Part 2 of their AI fluency series, Alane Boyd and Micah Johnson of Automate Your Agency break down the six organizational milestones that separate companies who are dabbling from those who are actually building. Whether you're a team of five or fifty, these milestones give you a concrete framework to know where you stand, and what to work on next.

Milestone 1: Create an Acceptable Use AI Policy

Here's a sobering stat: only 30% of companies currently using AI have an acceptable use policy in place. That means the vast majority are letting team members use AI tools with zero guidance on what's approved, what data can be inputted, and how the organization expects those tools to be used.

Your policy doesn't need to be perfect on day one. Start with the basics: what platforms are approved, what information can and cannot be entered, and whether usage is encouraged or restricted in specific contexts. Then iterate. Alane and Micah's own AI policy has evolved significantly over four years as platforms changed and their usage deepened. The key is that it exists, and that it grows with you.

Milestone 2: Build an AI Committee and Identify Your Champions

Rolling AI out to an entire organization without champions is a recipe for chaos. Before you scale anything, you need a small group of people, an AI committee, who are actively testing tools, surfacing wins and roadblocks, and making decisions about what should be standardized across the company.

This isn't just about having enthusiastic early adopters. It's about creating a structured process so that AI decisions don't happen in silos. Champions help you identify what your company's specific use cases are, what the standard workflows look like, and how to bring the rest of the team along, especially the skeptics.

Micah's approach to winning over skeptics? Show them how AI benefits them personally. Until someone sees or feels a direct benefit to their own work, there's no conversation to be had. Once they do, like saving 45 minutes with a single Skill that took five minutes to build, they convert. Every time.

Milestone 3: Measure Your Team's AI Fluency

You can't improve what you can't measure. And yet most organizations have no idea where their team members actually stand when it comes to AI competence.

This milestone is about establishing a baseline. Using the individual AI fluency milestones from Part 1 (check out EP 106 for those), you can track where each person on your team is, identify who needs support, and build a plan to move everyone forward. Your AI champions play a huge role here; they're often the ones best positioned to help assess and develop fluency across the team.

Milestone 4: Create a System for Sharing Wins and Losses

If your team's AI discoveries are staying in individual inboxes or forgotten in one-off conversations, you're leaving enormous value on the table. Organizational AI fluency requires a shared system, a way for anyone to surface what's working, what isn't, and what the team should know.

For Alane and Micah, this looks like a dedicated Claude Usage channel in Slack, plus bi-weekly team training calls that include a "what cool thing did you use AI for this week?" segment. Other organizations do it through department meetings, leadership check-ins, or dedicated breakout sessions to work through specific workflow challenges together.

The format doesn't matter as much as the habit. Build the system, make it a regular part of how your team communicates, and watch the collective knowledge compound.

Milestone 5: Get Your Data Clean

This is the one nobody wants to hear, but it's non-negotiable.

AI is only as good as the data it has access to. And no, AI is not going to magically fix your disorganized folder system, clean up your naming conventions, or make sense of years of messy files. It can support the process of getting to clean data, but the work is still yours to do.

Micah's recommendation: stop trying to migrate the chaos. Start fresh. Build a clean knowledge base from scratch, curate only the data and context you want AI to access, and leave the old mess where it is. It's the same approach that worked when Alane and Micah moved from Basecamp to Asana; start new, and move things over only as needed.

Clean naming conventions, centralized storage, and intentional data curation aren't glamorous. But they're the foundation that everything else is built on.

Milestone 6: Document Your Workflows

Every automation, every Skill, every agent your team builds needs to be documented in plain language that anyone in the company can understand, regardless of their technical background.

Why does this matter so much? Because when something breaks, you don't want the only solution to be "find the person who built it." And when that person leaves the company, you definitely don't want to be starting from scratch.

Good workflow documentation answers three questions: What does this do? How does it work? What should happen if it doesn't fire?

The good news: this used to take hours of manual work. Now, Alane and Micah's team uses a Skill in Cowork to feed in one or more workflows and generate a full "how it works" document (complete with a high-level diagram) in seconds. It's one of the things they covered in their latest AI Agent Cohort, and it's a game-changer for teams that have been putting off documentation because it felt too burdensome.

The Bigger Picture: AI Is Never "Done"

Perhaps the most important takeaway from this episode is the one Alane put so clearly: AI is not a "check the box and move on" situation. It's an evolution.

Two things keep this true indefinitely. First, the tools themselves are changing rapidly. Cowork in mid-2026 looks dramatically different from Cowork at launch in January 2026. Six months from now, it will look different again. Your team's fluency has to keep pace with that.

Second, turnover means you're always onboarding someone who is starting at zero. Building AI fluency into your new hire onboarding and assuming that even experienced hires need to learn how your company uses these tools, is what separates organizations that scale their AI knowledge from those that constantly rebuild it.

Hit these six milestones, and you won't just be a company that uses AI. You'll be a company that grows with it.

Want to go deeper? Listen to EP 107 of the Automate Your Agency podcast wherever you get your podcasts. And if you're ready to build AI agents for your business, the next AI Agent Cohort with Micah Johnson starts July 20th.

Most companies say they're using AI, but only 30% have even created an acceptable use policy. Alane Boyd and Micah Johnson are back with Part 2 of their AI fluency series, and this time they're going deep on the organizational level: the six milestones that separate companies dabbling with AI from those actually building with it.

If your team is experimenting with AI tools but you have no committee, no shared systems, and no way to measure who knows what you're building on sand. This episode gives you the framework to change that, no matter where your organization is starting from.

In this episode, you'll learn:

• How to build an acceptable use AI policy that evolves as your tools and workflows evolve
• Why AI champions matter and how to identify and leverage the right people on your team
• How to measure your team's AI fluency using milestone tracking across roles
• The system for sharing wins and roadblocks so AI knowledge doesn't stay siloed
• Why clean data is non-negotiable, and why AI can't fix your messy folder system for you
• How to document your workflows so anyone can understand them, no technical background required

If you're serious about building an organization that doesn't just use AI but grows with it, this episode gives you the exact roadmap. Press play. Your team will thank you.