Small team, Shopify eating the week: the AI that helps.

Ask a small ecommerce team where the week went and you get a list of small jobs that never show up on a plan. This post collects the published numbers on that list: how many hours UK owners say admin takes, how many small firms use AI, how many apps a store carries, and what a chat widget on a storefront does and does not do. Where a source states its method, the method is given. Where it does not, that is said too.

Where owners say the hours go

The most recent UK figure comes from American Express and Small Business Saturday UK research, reported by Startups Magazine in July 2026. It surveyed 1,000 UK micro, small and medium sized business owners and found an average of 11 hours a week on administrative or finance related tasks. Time on sales and growth came to just over half of that, 3.6 days a month. The figures are self reported estimates, and the sample is mixed across sectors, so none of it is specific to a Shopify store.

Sage's research, reported by SmallBusiness.co.uk in an article updated in January 2024, gives a different slice: an average of 120 working hours a year on administrative tasks, with 5.6% of staff time in each UK small business going to back office work. Accounting administration was the largest part, above 20% of all admin time. The article does not state the sample size or the year of fieldwork, so treat the figure as indicative.

The gap between them is large. Sage's 120 hours a year works out at a little over two hours a week, while the American Express research gives 11. Part of the difference is scope: the 11 hours cover finance and administration together and come from owners estimating their own week, whereas the Sage figure is back office time averaged across staff. Neither is wrong. Both are a reminder that admin hours depend on what the question counts, so count your own before comparing yourself with a headline.

The two surveys measure different things and should not be added together. What they share is the direction. Owners put a day or more a week on work that is not selling, and finance is only part of it. For an online store the rest is the admin around the catalogue and the inbox, which neither survey counts separately. We found no published survey with a stated method that splits a Shopify team's week into product listing time and customer service time, so the next two sections use platform data from tools merchants already run.

• 11 hours

  a week on admin or finance tasks, reported by UK business owners

  American Express and Small Business Saturday UK, 2026

• 28%

  of UK businesses with 0 to 9 employees using at least one AI technology

  ONS BICS wave 159, June 2026

• 6

  apps installed on the average Shopify store

  Shopify Partners blog

What the inbox is made of

Gorgias, the helpdesk vendor, publishes platform data from ecommerce merchants. Its customer service benchmarks say up to 30% of incoming tickets are shipping status requests. Its guide to auditing ticket volume names returns and exchanges, and sizing and product questions, as the other large groups, and tells teams to take the last 30 days of conversations across every channel and sort them into shipping, returns, product questions and account issues.

For a small team the useful part is the reading of each group. A pile of sizing questions means the product page lacks something, and a pile of returns questions points at a policy the customer finds confusing. Those are catalogue and policy fixes that sit upstream of the inbox, which is why the hours do not show up where you expect them. The week goes on the same answer typed again and on the page that would have made the question unnecessary.

Speed is the other number the helpdesk gets judged on. Gorgias puts the all industry median first response time at 6.3 hours for stores at the 10 million dollar sales band, with a spread from 1.6 hours for hardware to 8.8 hours for apparel in its March 2026 data. A small team answering in a working day is not an outlier in fashion, and would be in hardware. The takeaway is to compare against sellers of your own kind of product, and to decide your target before a chat window sets it for you.

Gorgias also reports on handoffs, the moment a customer is moved from one channel to another. Its Ecom Lab data shows a contact form handoff with a 36 hour median resolution time and 42% abandonment, an email handoff at 32 hours and 30%, and a chat handoff at 8 hours and 13%. Chat resolves faster in that dataset, which is a fair argument for having it. It is also an argument about speed on a channel, since a chat window does not change what the customer asks.

What the catalogue chores look like in data

MESA, which makes a workflow tool for Shopify merchants, analysed 1,374 workflows built by merchants in the first quarter of 2026 and published the method. Tagging ranked as the primary automation focus, with one in four merchants requesting tagging automation. Order automation, inventory, customer experience and fulfilment followed. Only 50.1% of the workflows created were actually shipped, which suggests a lot of small teams start an automation and never finish it.

The numbers describe merchants who already use a workflow tool, so they lean toward the technical end of the market. They are still useful as a picture of what a small team tries to take off its plate: tags that drift, orders that need a status change, stock that needs a threshold. Tagging being first fits the Gorgias reading above, since consistent tags are also what let a team sort the inbox by cause, and Shopify's own guide recommends tags that record why a problem happened.

MESA also reports that workflows using AI shipped at a rate of 70.3%, about 20 points above the baseline. That is a vendor's own analysis of its own users, with no control group, so it shows that people finish AI assisted workflows more often than others, not that AI makes the work better. One plausible reading, which the data cannot confirm, is that drafting help lowers the cost of starting a task.

How many apps a store already carries

Shopify's own partner guide to building apps states that the average Shopify merchant has six apps installed. The guide does not give a date or a method for the figure. It does explain the size of the pile a small team is managing: six separate sets of settings, six permission prompts that were once approved, and six places where something can fail when a theme changes.

Apps are also where the chat widget usually comes from. Shopify defines live chat as a messaging tool that lets customers have a real time conversation with a business through a chat window on a website, app or social channel. Some setups route every message to a staff member. Others use AI to answer common questions and pass the conversation to a person when a customer asks, or when the case needs human judgement.

What a storefront chat widget does and does not do

A widget does one job well: it gives a visitor a place to ask a question at the moment they have it. Shopify's guide says an experienced agent can handle four to six chats at once, compared with one phone call at a time. For a small team the same guide suggests using analytics to find the busiest hours and focusing availability on high intent pages such as product and checkout pages, instead of staffing chat all day.

There are three things a widget does not do. It does not write or fix the product descriptions that generated the question. It does not tidy the tags, vendors and product types that make the catalogue hard to search. And it adds an inbox of its own, so a team of three now has a chat queue, an email queue and the admin. A small team gets value from chat when the questions are already well answered elsewhere, and loses time when chat becomes one more place to type the same sizing answer.

Gorgias's figure that up to 30% of tickets are shipping status requests sits oddly beside the idea of chat as the fix. A customer who asks where an order is wants a status, and a widget can answer that if it can see the order. Whether it can depends on the integration, and checking that is the first test of any widget.

How many small firms already use AI

The Office for National Statistics publishes the best UK series, from its Business Insights and Conditions Survey. The release of 20 July 2026 covers wave 159, run from 5 to 28 June 2026, with 38,637 participating businesses and a 26.7% response rate. It excludes agriculture, finance, public administration and a few other sectors. Among businesses with 10 or more employees, AI use rose from around 12% to around 35% since late 2023. Among businesses with 0 to 9 employees, 28% reported using at least one AI technology, against 49% of those with 250 or more.

The gap between a five person company and a large one is 21 percentage points, narrower than most people expect. The most common use the ONS reports is improving operations, with close to 60% of businesses using AI for that purpose. The release does not break out ecommerce teams or customer service use for firms under 10, so any claim about a Shopify team in particular goes beyond the data.

Reading these together gives a practical test for a small team. List the three jobs that take the most hours: from the sources above, expect finance, the inbox and catalogue upkeep. Then ask of any tool whether it removes one of those jobs, or only adds a channel. Count the apps you already run, because six is the average and each one is a configuration to maintain.

The ONS also asks what holds adoption back. Among businesses with 100 to 249 employees, around 18% say insufficient expertise has delayed adoption. Cost is a constraint in every size band, reported by around 7% to 14% of businesses. And 41% of businesses with 10 or more employees report no barriers at all. Industry matters more than most of these: almost three fifths, 58%, of businesses in information and communication use AI, against 13% in construction. The release gives no separate figure for retail or ecommerce, so a store should not assume it sits near either end.

One more number frames the speed side of the inbox. Shopify's customer service guide cites research that 80% of customers now expect a response within 24 hours and 37% within an hour. It advises sellers to track the support volume ratio, the percentage of orders that result in a support request, and to use the feedback to remove confusion. A small team can compute that ratio on a Friday afternoon from the helpdesk and the order count, and it shows whether the inbox is growing faster than sales.

A week audit is cheap and does not need software. For five working days, write down each task as it starts, with a rough time and one of four labels: catalogue, inbox, orders, other. Most small teams that do this find the catalogue and inbox labels carry more minutes than anyone guessed, and that a handful of repeated tasks account for most of them. That list is the brief for any tool, and it keeps the buying decision tied to hours you counted rather than hours a vendor promised.

Where BYOM fits

What helps is drafting the work and stopping before publish. That is Kina's job in BYOM. Kina reads the Shopify catalogue and stock, and the helpdesk you already use. It does not publish on its own.