How AI is really changing life inside a small multichannel pet brand

Most AI and retail coverage is about enterprise tools. Or a shopping future that’s always five years away.

Nobody talks about the small startup selling the same product on Shopify, Amazon, eBay and TikTok at the same time. Four plates to spin, four sets of rules.

I’ve seen this from both sides. After 17 years in Amazon and ecommerce, I built and exited my own marketplace agency. I worked with Reckitt, Mars Petcare, Imperial Brands, Unilever, Brabantia and Polaroid, helping them scale on Amazon and other ecommerce channels. More recently I studied AI for Business Innovation at Imperial College, and I advise DragonFly AI and Clarisix.

Now I run The Catnip Queen, a premium cat wellness brand. It’s small and resource constrained, and it’s where I find out whether my advice to clients holds up when it’s my own money.

Research is where AI earns its keep

For me it’s the most useful application by some distance.

Proper market research used to mean five to seven days of spreadsheet work, or tens of thousands in agency fees. Customer segmentation, competitor positioning, category gaps, what buyers worry about before they click.

Now a solo founder can get answers in hours.

They’re not perfect answers. AI hallucinates and carries bias, so nothing goes into the plan on trust. But the speed and quality are good enough to make decisions with, which is what a small brand needs.

I also use MCP connectors to analyse advertising data quickly. Someone still has to map the data points properly and spot when the numbers don’t look right, that nous isn’t made redundant by AI.

Everyone gets the same data from the same tools. What you do with it is still yours, and I’ve never seen a human shaping the strategy matter more.

Amazon: writing for a shopper that isn’t human

Listings used to be about keywords, find what people search, put it in the title and bullets, watch the rankings move.That still matters of course but Amazon now has COSMO and Rufus, AI shopping assistants that interpret what a product is for, who it suits and whether it answers the shopper’s question.

That’s a different skill from ranking on Google, and most small sellers haven’t clocked it yet. Keyword stuffing doesn’t help a system trying to understand your product in context.

“AI doesn’t just find products anymore, it forms opinions about brands. Launching The Catnip Queen has taught me that being ‘prompt worthy’ matters just as much as being search optimised.

Small brands that treat their identity as an afterthought will find themselves invisible to the tools their customers now trust most.”

The fix is simpler than people expect. Say what the product does, who it’s for and why it’s different, in plain language.

Shopify: fast, but only with a voice

Copywriting and blog production that used to take weeks now takes days. I can also fix Shopify theme SEO issues without a developer, through a direct connection from Claude.

The trap is that AI content without a strict brand voice is instantly recognisable and instantly forgettable. Polite, tidy, says nothing.

For The Catnip Queen the work happens before the first prompt. Tone of voice rules, what we say, what we never say, how we handle claims. AI then works inside those guardrails. It also plays nicely with AI discovery. The right plugins get a Shopify site optimised for AEO very quickly.

eBay and TikTok: same cat, different page

Both have leaned into AI assisted search and recommendation, and both reward different listing structures. So each channel needs its own version of the product page, with different titles, emphasis and story.

A few years ago that would have been overkill for a small brand. Now it’s standard, and AI makes it manageable because the variations take minutes.

But speed doesn’t replace knowing the channel and its algorithm. I’ve noticed that TikTok rewards being a real human, too.

The backend

Some of the biggest shifts are the ones customers never see. Ad copy testing, Google RSA campaigns built to the exact character count, keyword gap analysis. All of it used to sit behind an agency retainer, a lot of it can now be done in house.

But I’m not saying agencies are dead, I ran one. Good ones sell judgement, strategy and accountability.

But paying a retainer for execution a Founder can do in an afternoon is a harder sell. That frees budget for product and creative, which is what excites me most.

Small versus big

“Also in the world of AI solo entrepreneurs can build a brand including, sourcing, forecasting, marketing and advertising all aided by AI. This means the playing field is being levelled dramatically for SMBs vs CPG brands.”

I’ve worked with the big CPG names. Brilliant people, serious budgets, serious scale. What they often lack is speed, because decisions pass through layers of C suite, PR and legal.

A small brand with clear positioning can test, learn and change direction in a week. Being small used to make that a side benefit. With AI it’s a real edge.

Treat an AI like an MBA Graduate

I have trained many graduates and apprentices over the years. Their intellect and talent sometimes astound me, but as an owner manager you would also want them fully trained on SOPs and enlightened to the ‘real world’ of business outside the academic world.

This approach is also true to an AI, think of your LLM as a super intelligent MBA graduate and coach them and check their work (like you would a human).

An AI can make mistakes in the same way a Master’s Graduate does, a solid conclusion to a project may be incorrect, not by their fault, but mainly yours or the company’s not providing the training (Project set up, skills or context) or poor prompts not providing the full picture.

My point is take time to train an LLM and provide accurate context for data points and you will get sound outcomes.

What I’d tell a founder starting out

AI commoditises execution, not judgement or strategy. Anyone can write a listing or build a campaign now so those stop being the differentiator. What’s left is knowing your customer, knowing why your brand exists and having a point of view.

If I were starting today, I’d do three things:

• Write your identity down first. Voice, positioning, what you stand for. AI can only reflect what you give it.

• Treat every platform as its own game. Same product, different rules, different judges.

• Use AI for the heavy lifting. Keep the strategy in your own head.

Know who you are before you open the tool.

For more startup news, check out the other articles on the website, and subscribe to the magazine for free. Listen to The Cereal Entrepreneur podcast for more interviews with entrepreneurs and big-hitters in the startup ecosystem.