Stage 1 to Stage 2: The Quick-Start Guide for Moving from Scattered to Standardized AI

If your team is using AI, but nobody decided to use AI, you're at Stage 1.

That's not an insult. It's where most businesses are. ChatGPT launched two years ago and people started using it on their own. Some of it has been genuinely useful. A lot of it has been random. Almost none of it was deliberate.

Stage 2 is the first real decision point. You stop letting AI happen to your business and start making intentional choices about how it gets used.

Here's what that looks like in practice.

---

What's Actually Different at Stage 2

Stage 2 doesn't mean everyone on your team has become an AI expert. It means the business has made three decisions:

1. Which tools are we using (and which ones we're not)
2. What those tools are for — and what they're not for
3. How we handle data, confidentiality, and outputs

That's it. You're not building automation or running agents. You're just getting the team on the same page about a set of tools you've actually vetted and approved.

The difference from Stage 1 is that this is a business decision, not an individual one.

---

Step 1: Do a Quick Audit

Before you standardize anything, find out what's already in use.

Ask your team a simple question: "What AI tools are you currently using, and what are you using them for?" You might be surprised. Some of it will be useful and repeatable. Some of it will raise questions about where company data is going.

Give yourself one week to gather this information. You don't need a formal survey. A team meeting or a quick email works fine.

What you're looking for:

• Which tools are people actually using
• Which tasks are getting done with AI
• Whether any sensitive data (client info, financials, employee records) is being fed into these tools

---

Step 2: Pick One Primary Tool and Commit to It

The biggest mistake at Stage 1-to-2 transition is trying to standardize five tools at once. Pick one. Get your team comfortable with it. Add tools later.

For most small and mid-size businesses, the right starting point is a capable general-purpose AI assistant. ChatGPT, Claude, Gemini — there are good options at low cost. The specific tool matters less than the commitment to use it consistently.

Pick the one your team is already gravitating toward. Consistency beats optimization at this stage.

---

Step 3: Set Three Simple Rules

You don't need a 20-page AI policy. You need three rules your team can remember without looking them up.

Here's a starting point:

Rule 1: No client data in AI prompts without explicit approval. Names, contracts, financials, contact information — none of it goes into a public AI tool unless you've made a decision that it's acceptable.

Rule 2: AI outputs get a human review before they go anywhere. An AI-written email still gets read before it's sent. An AI-generated report still gets sanity-checked. The human is responsible for the output.

Rule 3: If you find a use that works well, share it. The fastest way to move through Stage 2 is for people to share what's working. A quick Slack message or a 5-minute team standup item — "I used AI to do X and it saved me an hour" — builds organizational knowledge faster than any training program.

---

Step 4: Build One Repeatable Use Case

The gap between Stage 2 and Stage 3 is whether AI is built into your workflows or just available to your team. You don't need to close that gap yet. But you do need to identify it.

Before you leave Stage 2, find one use case that's genuinely repeatable and valuable, and document it.

Examples:

• AI-assisted first draft of all client proposals
• Weekly summary of team status updates drafted by AI, reviewed and sent by a manager
• Meeting notes summarized and formatted by AI immediately after every call
• Customer emails drafted by AI, reviewed and sent by the service team

Document it simply: what triggers the task, what prompt or instructions the team uses, what they check before sending or saving. One page. No overengineering.

That document is the seed of Stage 3.

---

What Stage 2 Actually Looks Like When You're There

Your team is using one or two approved tools consistently. You have basic rules about data and review. At least one workflow uses AI in a way that's repeatable and has been written down.

You haven't automated anything yet. You haven't changed how your business is structured. But you've stopped improvising and started building.

That's Stage 2. It's not glamorous, but it's the foundation everything else sits on.

---

How Long Does This Take?

For most businesses: 2 to 4 weeks if you're focused. The audit takes a few days. The tool decision is usually fast — most teams already have a preference. The rules take one session. The repeatable use case takes another week or two to document.

This isn't a multi-quarter project. It's a focused sprint of decisions.

---

Not Sure Which Stage You're At?

Take the 8-question AI Maturity Assessment at lariatco.com/assessment. It takes 2 minutes and gives you a specific stage result with next steps based on where your business actually stands.

Take the Assessment at lariatco.com/assessment