How 3 small business owners are — and aren’t — using AI to nail their marketing campaigns

Marketing is an integral part of running a small business, but building a successful strategy can be time-consuming for owners. Can AI tools geared toward marketing help?

Meta thinks so. In August, it announced new AI features for small business owners, designed to help audit organic Facebook and Instagram posts and ads, measure businesses’ performance against similar brands, and offer strategy suggestions to improve marketing efforts. Solo business owners have also tapped generative AI tools like ChatGPT, Claude, and Gemini to assist with time-intensive marketing tasks like customer outreach and content ideation.

We asked three small business owners about how they think through and implement marketing strategies in the AI age.

The following has been edited for length and clarity.

How do AI-assisted tools fit into your overall marketing strategy?

Andrea Garcia Fry, cofounder at baby and children’s clothing brand Arquia: Effective paid media requires an understanding of your brand positioning, tone of voice, business goals, and customer journey.

I prefer using ChatGPT or Claude because I can provide fuller context and use them as thought partners. Meta’s tool will be valuable if it genuinely learns the business rather than simply recommending more automation or spending.

Catalina Parker, founder of the coaching business Relatable Nonprofit: I’ve tried implementing recommendations directly from Meta Ads Manager, and many of them either didn’t work or weren’t right for my business. Meta benefits when advertisers spend more money on its platform.

I’d rather have a tool that helps me independently analyze the data across my entire marketing ecosystem instead of relying solely on Meta’s interpretation of what I should do next.

Bonnie Gurry, CEO at climate-friendly investing platform GreenPortfolio: Tools like Meta AI can be great operational time-savers. It’s like we now have a junior analyst to run quick diagnostics, spot creative trends, and crank out copy variations in seconds. Similar tools from other ad networks have saved us hours of manual dashboard digging so we can focus on core growth strategy.

But you can’t let the platform mark its own homework. It’s a great assistant for speed, but I wouldn’t cede capital allocation or brand positioning decisions to it without validating that its recommendations align with business objectives.

What’s your biggest challenge when it comes to paid marketing?

Garcia Fry: Paid marketing can become expensive very quickly. Meta needs time and budget to learn who the right customer is, which means a meaningful portion of your investment can be spent during that learning process without producing the results you expected. After that, success isn’t guaranteed.

Finding the right customer on Meta has been difficult for us, especially with a limited budget. For Arquia, strengthening our brand foundations and organic marketing has brought us better results than our paid campaigns.

Parker: Understanding performance across the entire funnel. A lot of people talk about paid ads as the way to scale growth. I learned the hard way that simply putting more money into a campaign because the leads are cheap is not necessarily the answer.

You have to track what happens after someone sees the ad. Are they clicking? Opting into the lead magnet? Engaging with your content? Booking a call? Eventually buying?

With AI-generated content, bots, and increasing skepticism online, trust feels harder to earn. I’m trying to combine paid acquisition with something much more old-fashioned: consistently nurturing people and building real relationships over time.

Gurry: Knowing why an ad worked. Was it the opening hook, the visual, or just a warm retargeting pool?

Isolating those creative variables to figure out what actually drove performance, without burning through capital, is a constant balancing act.

What is a tool or strategy helping you succeed?

Garcia Fry: When I run Meta ads, I start with what works organically. If a piece of content resonates with our audience, I may boost it or adapt it into a paid campaign.

I also use ChatGPT or Claude to audit the setup, interpret the data, and identify possible tests, but I still make the final decisions.

Parker: I’ve learned not to assume that every recommendation from an expert will work for my business. Testing has been more valuable: knowing which numbers matter, tracking them consistently, testing changes, and using the results to decide what to do next.

One of my best-performing campaigns has been running for more than two years with very few changes. It’s a simple text-based ad, nothing fancy. It continues to perform despite all the advice about constantly producing new formats.

Gurry: We don’t need fancy, over-complicated tools. We aim for a clear, back-to-basics look at which ads bring in real clients over time.

Quick clicks can burn through a marketing budget with little return on investment if you are not careful. Clicks are great, but not if they aren’t reaching your potential customer.

And don’t outsource strategy to an AI tool — especially one developed by the network that’s selling you ads.