You may be tired of hearing how AI is going to affect how your business works, whether you’re planning for that change or not. The truth is you’re probably not thinking big enough.
Key Takeaways
- You’re probably not thinking big enough about AI. Asking it one question at a time is only the first stage. The bigger gains come when you give it context about your business and a clear process to work in.
- Saving time is a start. Decide what the business needs to improve. Even among marketers already enthusiastic about AI, far fewer report significant, measurable results than report saving time.
- Start with the repeated work you don’t like doing. Document the steps and where a person needs to decide, then test AI there before you hand over anything bigger.
Most people are still using AI for a quick search or simple question. That’s a fine place to start, but it’s only the first step in what AI can do for your business.
You’re Not Behind
With all the announcements, it can feel as if everybody else has already figured this out. Adoption is still not as widespread as the headlines suggest. As recently as April, only about 2 percent of US households were paying for an AI service, and that number has grown.
Paying isn’t the same as using AI, but the AI model capability on paid plans is better than free plans. It gives you some perspective on how early this still is. You have room to learn. You also have a reason to start now.
Saving Time Is Only the Start
There’s also a difference between using AI and getting a business result. In Social Media Examiner’s 2026 AI Marketing Industry Report, 86 percent of respondents said AI saved them time, but only 30 percent reported significant, measurable results. And that’s a self-selected group of marketers who are mostly tech-savvy and already interested in AI.
Saving time is helpful. But what are you doing with that time, and is the business better for it? If AI helps you produce twice as many posts about a service you don’t really want to grow, you’ve made the wrong work faster. Decide what needs to improve before adding more output.
As Ethan Mollick puts it, “When content can be generated instantly, the constraint isn’t production, it’s knowing what will actually matter to people.”
The Five Stages of Using AI
That’s where thinking beyond “help me write this email” matters. I think about the evolution of AI use in five stages.
[chart: five stages of AI use]
Stage 1: Task Prompter
Simple one-off prompts. You ask AI to fix an email or answer a search question. This is where most people start.
Stage 2: Brainstorming Buddy
You work back and forth on an idea instead of accepting the first answer. That could be researching a product or a solution to a problem. You explain your concerns and requirements, it suggests ways to address them, and you challenge what doesn’t fit. You’re using the conversation to think through the work.
Stage 3: Personalized Assistant
You give AI context about your business: the services you want to grow, the customers you work best with, your way of explaining things, and examples of good work. You save that context in a tool that can reuse it, so every task doesn’t begin with a blank page. If you want your quotes in a certain format but personalized for each client, the AI now knows enough to do a good job without starting over each time.
Stage 4: Workflow Partner
AI becomes part of a repeatable process. You take the notes and approved photos from a finished project, turn them into a useful website story, then adapt that story for your email and social channels. The instructions and review steps are already set up, and AI can help with the steps around the writing too, like flagging missing information and tracking what still needs approval.
Stage 5: Business Agent
Agents can handle parts of a process, or all of it, on their own. In a team of agents, each one handles part of a job and passes the work along under your rules. One could organize incoming inquiries, another gather the relevant service information, and another prepare a follow-up for your team to approve. They need clear responsibilities and access to the right information.
That also means deciding where a person takes over. An agent may help prepare a response, but promising a project price or a completion date requires authority and information you may not have given it.
You don’t have to jump straight to a team of agents. But if all you ever do is ask isolated questions, you won’t discover how AI could help with the work that repeatedly takes your time.
Start With the Repeated Work You Don’t Like
So where do you look for that work? Start with something that happens often and has understandable steps. I like to start with the repeatable things I don’t like to do, rather than the parts of the work I really enjoy.
Maybe every estimate requires collecting the same information, or your team repeatedly answers the same service questions. Write down how you handle it now, including the moments that make you say, “This needs my attention.” That gives you a place to test AI’s help without handing over the whole job.
And don’t automatically assume that automating what you do now is the right approach. You may find a more efficient workflow that lets everyone be more productive and improves the service you provide.
Lead the Change in Your Business
Your employees are going to use AI, too. Have you asked what they’re trying? Are you giving them direction about the information they can use and the work that needs review? You need to lead that learning.
The change also reaches you from outside. AI agents will start contacting businesses on behalf of customers, even if you never open an AI tool yourself. So this belongs in your business planning. It’s a bigger decision than choosing an app you happen to like.
That’s why we work on strategy and repeatable processes in our Marketing Accelerator Program. We help you get clear about what you want to accomplish, how the work should happen, and where AI helps you get your marketing done. Random acts of AI don’t fix your random acts of marketing.
Action Steps This Week
Three things to do this week:
- Identify which stage describes most of your AI use today and choose one next step beyond it.
- Pick one repeated business task and document its inputs, steps, and the decisions that need a person, then use that to run a small AI-assisted test.
- Review the result against a business outcome, like time saved after checking, fewer missed follow-ups, or better-qualified inquiries, and use what you learn to decide whether to expand it.
Your business is going to change as your team, your customers, and the tools around you use more AI. Start building the understanding to lead that change now. You can learn through a manageable piece of work while thinking much further ahead about how you want the business to operate.
Do you prefer to listen in? Here’s our podcast:
Links in this episode: a16z: State of Markets II
