- What people think they're buying versus what they get
- The honest hour breakdown from a real 20-hour month
- Where the hours go on customer service specifically
- The part that makes clients uncomfortable
- A realistic first 30 days, step by step
- What this costs when you translate hours into invoices
- The hours nobody bills for but every good consultant does anyway
- How to tell if your hours are being spent well
- Frequently asked questions
- Useful references
The short version: most of an AI consultant's hours on an ecommerce account go to unglamorous work like fixing product data, untangling broken automations, and training staff to use the tools they already pay for, not building flashy new chatbots. If a consultant's proposal is mostly "implementation" and light on "audit and cleanup", ask harder questions before you sign anything.
What people think they're buying versus what they get
When an ecommerce founder hires an AI consultant, they usually picture something dramatic: a chatbot that closes sales while they sleep, product descriptions writing themselves, ad spend optimising itself into a lower CPA. That stuff does happen. But it's maybe a third of the actual working hours, if that.
I ran a retainer for a home fragrance brand doing around £160,000 a month in revenue across Shopify and Amazon. The founder's brief was "we want AI in the business." Fine, but that sentence has no shape to it. The first thing I did wasn't build anything. It was spend nine hours going through their tech stack: Klaviyo, Gorgias, a Zapier account nobody remembered setting up, a Shopify app for reviews that had been silently double-charging them for eleven months. Nine hours, no AI touched, and it saved them £2,200 a year on the spot from that one billing error alone.
The honest hour breakdown from a real 20-hour month
On a typical 20-hour monthly retainer with an ecommerce client, here's roughly how the time splits, based on the last four accounts I've run:
- 6 to 8 hours on data and systems work: cleaning product feeds, fixing customer tags, checking that the "AI" tool is even getting accurate inputs
- 4 to 5 hours on building or configuring something new: a chatbot flow, an email segment, an automated reorder trigger
- 3 to 4 hours on testing and fixing what was built, because the first version almost never works cleanly on ecommerce data with real customer edge cases
- 2 to 3 hours on training staff, usually the same two people, on how to use what's been set up
- 1 to 2 hours on reporting back to the owner in language that isn't full of jargon
Notice what's missing: strategy sessions where everyone drinks coffee and talks big vision. Those happen once at the start and maybe once a quarter. The rest of it is closer to plumbing than to consulting in the way most people picture consulting.
Where the hours go on customer service specifically
Live chat and chatbot work eats more time than founders expect, and it's worth breaking down separately because it's the most requested AI project on ecommerce accounts. A brand I worked with wanted an AI chatbot handling return requests. Sounds simple. It took 14 hours across three weeks: mapping their actual return policy exceptions (there were nine, none written down anywhere), writing the conversation flows, testing against real past tickets pulled from Gorgias, and then rewriting half of it because the tone came out sounding like a legal disclaimer rather than a person.
If you're at the stage of picking the underlying platform before any of that flow work starts, it's worth reading through how to choose live chat support software for ecommerce first, because the wrong platform choice adds hours to every project built on top of it later. Picking the tool badly at month one costs you at month six.
The part that makes clients uncomfortable
Here's the bit that doesn't get said much: a good chunk of an AI consultant's hours on ecommerce accounts go into undoing work another agency or a well-meaning in-house marketer already did badly. I've inherited three separate chatbot setups in the last two years that were technically "live" and technically "AI" but were answering customer questions with information that was eighteen months out of date, because nobody had connected them to a live product feed. Fixing that isn't glamorous and it isn't billable as "new AI strategy", it's billable as cleanup, and clients sometimes push back on paying for it because they think the first agency already did that job.
They didn't. Getting an AI tool switched on is not the same as getting it right, and most of the market treats those two things as identical because "right" takes longer and costs more to quote.
This is exactly why the questions you ask before signing matter more than the pitch deck. Anyone considering this kind of hire should go through the questions to ask before hiring an AI consultant before agreeing scope, because "how many of my hours will be maintenance versus new build" is a question almost nobody asks up front and it's the one that predicts whether you'll be happy in month four.
A realistic first 30 days, step by step
If you're starting an engagement from scratch, this is roughly what a competent first month looks like on an ecommerce account, hour by hour:
- Week 1 (6 to 8 hours): full audit of current stack, customer data quality check, list every tool that's currently switched on but unused
- Week 2 (5 to 6 hours): pick one high-friction area to fix first, usually customer service response time or abandoned cart recovery, and map the current process exactly as it runs today, mistakes included
- Week 3 (6 to 8 hours): build and test the fix in a sandbox or on a small percentage of traffic, not the whole store
- Week 4 (4 to 5 hours): roll out, train the team who'll own it day to day, set a check-in date 30 days out to see if the numbers moved
That's around 21 to 27 hours across the month, and notice it produces exactly one working improvement, not five. Consultants who promise to "transform your whole customer journey" in month one are either overselling or planning to bill you for hours they won't spend carefully.
What this costs when you translate hours into invoices
UK based AI consultants working with ecommerce brands typically charge somewhere between £90 and £250 an hour depending on experience and whether it's strategy, hands-on build, or training. A monthly retainer of 15 to 25 hours lands most small to mid sized ecommerce brands somewhere between £1,800 and £5,500 a month. A one-off audit, the kind that catches the billing errors and dead tools before anything new gets built, usually runs 8 to 12 hours and £900 to £2,500 depending on how tangled the existing stack is.
If you want a fuller breakdown of what shapes that pricing, there's a detailed guide on how much an AI consultant costs that goes through the variables that push the number up or down, things like how many sales channels you run and how much of your customer data lives in one place versus scattered across four apps.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Location matters less than people assume now that most of this work happens over video calls and shared dashboards, but if you'd rather work with someone regionally, there are consultants covering practical AI work for Manchester businesses and separate ones handling AI automation for Yorkshire businesses around Leeds, both worth a look if you want someone who can turn up in person occasionally rather than purely remote.
The hours nobody bills for but every good consultant does anyway
There's a category of time that rarely appears on an invoice: the hour spent explaining to a founder's spouse why the "AI" isn't going to replace the two customer service staff, or the twenty minutes reassuring a warehouse manager that the new automated reorder system isn't going to override his judgement on a supplier he doesn't trust. That's not billable strategy work by any normal definition, but skip it and the whole project stalls because the humans who have to live with the system day to day were never brought along. I've watched a technically perfect automation get quietly switched off within six weeks because nobody on the warehouse floor was consulted before it went live. The build was flawless. The rollout ignored people, and people win every time.
How to tell if your hours are being spent well
Ask for a simple monthly log, not a vague summary. It should show hours against specific tasks: "4 hours, cleaned product feed for chatbot accuracy" tells you more than "4 hours, AI strategy work." If a consultant can't or won't itemise it, that's worth a direct conversation, and if you're still deciding who to work with, comparing how different consultants explain their time is a decent filter on its own, whether that's someone local like the teams working with Melbourne businesses on AI automation or further afield in Los Angeles working with creators and ecommerce brands. The good ones all show their working.
I do this work for owner-led businesses: the process is set out at hiring an AI consultant and the service detail at AI automation consultant.
The sector-by-sector version of this is on AI consultant by industry.
If you run or market an online store, see my list of ecommerce blogs that accept guest posts.
Frequently asked questions
How many hours a month does an ecommerce brand need from an AI consultant?
Most small to mid sized ecommerce brands need somewhere between 15 and 25 hours a month once initial setup is done, with the first month often running higher, closer to 25 to 30 hours, because of the audit and cleanup work that has to happen before anything new gets built.
Why does so much time go into "cleanup" instead of new AI tools?
Because AI tools are only as good as the data and processes underneath them. A chatbot connected to an out of date product feed or a customer segment full of duplicate records will make confident, wrong decisions faster than a human would, so fixing the underlying mess has to come first even though it's less exciting to pay for.
Is it normal for a consultant to inherit and fix another agency's work?
Yes, and it happens more than people expect. A large share of ecommerce AI projects involve fixing a chatbot or automation that was switched on but never connected to live data, which is why it's worth asking any new consultant directly whether their quote includes auditing existing tools before building anything new.
What's a fair hourly rate for AI consulting work on an ecommerce store?
UK based AI consultants working with ecommerce brands generally charge between £90 and £250 an hour depending on whether the work is strategy, technical build, or staff training, with most retainers landing between £1,800 and £5,500 a month for 15 to 25 hours of work.
