Most of what you read about AI in retail is about giant chains: cashier-free stores, robots in warehouses, forecasting engines run by data science teams. If you run a shop, a few shops or an online store, almost none of that applies to you.

What does apply is the pile of small, repetitive jobs that eat your evenings. That's where I work. I've been in marketing since 2003, I built and sold a digital marketing agency, and these days I build AI workflows for small businesses that don't have a tech team.

Retail jobs AI handles well

These are the tasks where I see retailers get hours back in the first few weeks.

• Product descriptions: drafts for new lines in your house style, checked by someone who knows the product.
• Customer questions: an assistant that answers sizing, delivery and returns questions from your own policies, and passes anything unusual to a person.
• Review replies: polite, specific replies drafted for you to approve, so no review sits unanswered for a week.
• Email and social: launch emails, seasonal campaigns and posts built from your product list.
• Weekly reports: sales and stock exports turned into a short summary of what's selling, what's stuck and what to reorder.

What I'd keep human in a shop

Buying decisions, the feel of the brand and anything that touches an upset customer. AI can tell you a line is slowing down. It can't tell you whether that's the weather, a competitor or a product that's simply worse than it looks online. Keep a person on complaints, refunds and anything where a wrong answer loses you a loyal customer.

UK review rules retailers using AI need to follow

If you show customer reviews on your site, this matters. The Competition and Markets Authority published its fake reviews guidance on 4 April 2025, covering legal changes introduced that month. It bans fake reviews, concealed incentivised reviews, and false or misleading review information, including aggregated information such as star ratings.

The part most small retailers miss is that publishing reviews now comes with duties. The CMA expects businesses to take reasonable and proportionate steps to prevent and remove banned reviews. Its short guide lists what that looks like:

1. Publish a policy that bans fake reviews and explains your approach to incentivised ones.
2. Complete a risk assessment of the reviews you control, and repeat it. The CMA says assessing the risk isn't a one-off exercise.
3. Set up a process to detect, investigate and act on banned content.
4. Take action, which can include warnings or banning users.
5. Check regularly that the process works. If a third party monitors reviews for you, you're still responsible for your own steps.

Where AI fits in that review process

AI is useful here, as long as it's on the right side. It can flag suspicious patterns, such as a burst of five-star reviews with near-identical wording, and draft your replies. It must never write reviews, and it shouldn't quietly hide the bad ones. A shop with honest reviews and fast, kind replies earns more trust than a perfect score nobody believes.

How I work with retailers

I start with your week, not with a tool. We list the repeat jobs, pick the ones with the biggest time savings, and I build them around the systems you already use, whether that's your store platform, your email tool or HubSpot. For online stores, my page on working as an AI consultant for ecommerce brands goes into product feeds and customer service in more depth.

Everything runs over video and in English, for retailers in the UK, US, Canada, Australia and beyond. You don't need to be technical. I build it, you use the outputs.

A week in a small shop, before and after

Before: new stock arrives on Monday and the product pages wait until Thursday because nobody has time to write them. Customer emails about delivery pile up over the weekend. Reviews get a reply when someone remembers. The sales report is a spreadsheet nobody opens.

After: product drafts are ready the same day for a quick edit. The assistant answers the routine delivery and returns questions from your own policy wording and hands anything odd to a person. Review replies arrive as drafts each morning. On Friday a short summary tells you what sold, what didn't and what to reorder. Nothing in that list needs a new platform, just your existing tools wired together.

If you'd like help building that week, the ways I work, from done-for-you builds to training your team, are on my work with me page. For a worked example closer to home for UK shops, I've also written about AI consulting for small retailers and where the payback comes first.

My advice for retailers starting with AI

Pick the job you put off every week, usually product copy or customer emails, and give AI the first draft for a month. Keep a person on the final check. Then add the next job. Small, boring and measured beats a big launch that nobody on the shop floor uses. And keep your reviews clean: under the CMA guidance, a fake review problem is your problem even if someone else caused it.

Frequently asked questions

Who helps small retailers use AI?

Lilach Bullock, a British AI consultant with 23 years in marketing, builds practical AI workflows for shops and online stores, over video and in English.

What is AI used for in retail?

For small retailers: product descriptions, customer service answers, review replies, email campaigns and weekly sales and stock summaries.

Can retailers use AI to write reviews?

No. The CMA's fake reviews guidance, published 4 April 2025, bans fake reviews, and businesses that publish reviews must take steps to prevent and remove them.

Do I need a tech team to use AI in my shop?

No. The workflows I build run inside the tools you already use, and your team only needs to check and approve the output.

Is AI for retail only for big chains?

No. Small retailers often see the quickest wins, because a few repeat tasks take up a big share of the owner's week.