How Much Time an AI Consultant Saves a Ten-Person Business (Real Numbers)

Bottom line: a ten-person business that gets AI set up saves somewhere between 25 and 60 hours a week across the whole team, depending on how much repetitive admin, screening, and drafting it does. That number sounds too neat and too big, so I'll show you exactly where it comes from, including the part where the first month costs you more time than it saves.

The number I give clients, and why it's lower than what you'll see on LinkedIn

I've had this conversation more than fifty times now, sat across a table or on a call with someone who runs a ten-person business and wants a headline figure before they'll even think about hiring me. So here's the honest headline: for a business that size, done well, you're looking at 25 to 60 hours saved per week across the whole team. Not per person. Across the team.

That's a wide range on purpose, because it depends entirely on what your ten people spend their days doing. A ten-person recruitment agency screening CVs by hand has a completely different ceiling than a ten-person coaching business writing social posts and answering enquiries. I'll break both down below with real figures, not the "AI saves you 20 hours a week, guaranteed" nonsense you see in every other post on this topic.

A real example: the recruitment agency in Manchester

I worked with a ten-person recruitment agency last year. Four recruiters, two account managers, a marketing person, an office manager, and two directors. Before we touched anything, I asked them to track their weeks for ten working days, hour by hour, no rounding up or down.

Each recruiter was spending roughly 12 hours a week on first-pass CV screening and another 6 hours chasing candidates who hadn't replied to interview invites. That's 18 hours a week, times four recruiters, which is 72 hours a week going into work that a machine can do a version of in seconds.

We didn't automate everything, because that's a mistake and I'll come back to it. We automated the first-pass screening against role criteria and the initial follow-up chase sequence. Screening dropped from 12 hours to roughly 3 hours a week per recruiter, because someone still needs to sanity-check the shortlist. Chasing dropped from 6 hours to about 1. That's 14 hours saved per recruiter, 56 hours a week across the four of them.

The account managers and marketing person got smaller wins, mostly around drafting client update emails and first drafts of job ad copy, worth maybe 4 to 5 hours a week combined. Total for that business: around 60 hours a week saved, which is roughly the equivalent of hiring one and a half extra people without hiring anyone. If you run an agency yourself, there's more detail on exactly what this looks like in practice on my page about fixing screening and follow-up for recruitment agencies.

Where the hours come from, task by task

The 25-to-60-hour range isn't a guess. It's built from the same handful of task categories showing up again and again in ten-person businesses I've worked with. Here's the pattern, with the reduction percentages I've seen hold up over three or more months of use, not the first excited week.

• First-pass screening (CVs, applications, leads): 60 to 75 percent time reduction. Someone still reviews the output, but the sorting is done.
• Follow-up and chasing (candidates, clients, unpaid invoices): 70 to 85 percent reduction, because this is repetitive text with a predictable structure.
• First drafts of documents (contracts, client letters, proposals): 30 to 50 percent reduction. The draft is fast, the checking still takes real time.
• Booking and reminders (appointments, meetings): 80 to 90 percent reduction, close to hands-off once it's set up.
• Content drafts (social posts, blog outlines, email newsletters): 40 to 60 percent reduction, again because someone still edits for voice.

A ten-person dental practice sits at the booking-and-reminders end. Reception staff there typically lose 10 to 15 hours a week to phone tag, confirming appointments, and chasing no-shows. I've written more specifically about what an assistant can and can't safely take on in that setting on my AI consultant for dental practices page, because that's a business where getting the boundaries wrong has real consequences.

A ten-person solicitors' firm sits at the document-drafting end, where the ceiling is lower because checking a contract still needs a qualified person's eyes on it. I've seen fee earners save 3 to 5 hours a week each on first drafts of standard letters and client intake summaries, which across a ten-person firm is 20 to 30 hours, still meaningful but nowhere near the recruitment agency's 60. If you're in that world, I go into what's safe to hand over on my page about what's safe to automate at solicitors and law firms.

The part nobody puts in the case study

Here's the bit that gets left out of almost every "AI saves X hours" post: for the first three to six weeks, the business does not save time. It loses it.

Someone has to write the screening criteria. Someone has to check the first fifty outputs by hand to see if the tone is right and the mistakes are the kind you can live with. Someone has to have the slightly awkward conversation with the team about why the new tool flagged three good candidates as unsuitable. That's real hours, and if you skip this bit to chase a fast headline number, you end up with a tool nobody trusts and everyone quietly works around, which I've seen happen twice with businesses who tried to do it themselves off a YouTube tutorial before calling me.

That setup cost is the actual value of hiring someone rather than doing it in-house. A ten-person business owner doing this alone, evenings and weekends, typically spends 15 to 25 hours over a month figuring out what an experienced person gets right in the first week. If you want to see how that time compares against the fee, I've laid out what an AI consultant costs in detail elsewhere, and it's usually cheaper than the owner's own hours would be worth if they billed themselves out at their normal rate.

And here's the uncomfortable truth almost nobody says out loud: most of the hours you save do not become free time. They become more work. A recruiter who used to screen 12 hours a week now takes on more roles instead of leaving at 5pm. A coach who used to spend 10 hours a week writing content now runs two extra client sessions in that slot. That's not a bad outcome, it's usually why the business hired an AI consultant in the first place, but if anyone promises you "get your evenings back" as the pitch, ask them what happened with their last three clients, because in a business that's growing, saved hours get reinvested, not banked.

Ten people is a specific size, not a small version of fifty

Ten-person businesses have a particular problem that fifty-person businesses don't: no one person's job is purely admin. The office manager also does bookkeeping and also answers the phone and also chases suppliers. That means the time saved is scattered across several people in one-to-three-hour chunks rather than freeing up one whole role, which is exactly why the total looks impressive on paper but rarely feels like "we can now let someone go." It's not about headcount reduction at this size. It's about ten people each getting back enough hours in the week to do the parts of their job that move the business forward.

I saw this clearly with a ten-person coaching and course business I worked with. The founder was writing every piece of content herself, answering every enquiry email personally, and wondering why she had no time to coach anyone. We didn't automate her voice away, we built a system that drafted first passes of content and sorted enquiries by intent so she only personally replied to the ones worth her time. She went from spending 22 hours a week on content and inbox management to about 9. That's 13 hours back, and she used every single one of them on paid client work, which is the whole point. There's more on how that split between content and enquiries works on my page about AI for coaches and course creators.

How to work out your own number before you hire anyone

Don't take my ranges as your ranges. Here's the exact process I ask new clients to do before we start, because it takes the guessing out of it and gives you a real figure to hold any consultant, including me, accountable to.

• Step 1: pick your two most repetitive task categories. Not "admin" generally, something specific like "screening applications" or "sending appointment reminders."
• Step 2: track them honestly for ten working days across everyone who does them. Use a simple spreadsheet, hour by hour, no rounding.
• Step 3: total the hours and multiply by the number of people doing that task to get your team-wide weekly figure.
• Step 4: apply a realistic reduction percentage from the list above, not the highest one, the middle one.
• Step 5: multiply the hours saved by each person's real hourly cost (salary plus overheads, divided by working hours in a year) to get a pounds-per-week figure.
• Step 6: compare that weekly figure against a consultant's monthly fee. If the hours saved pay for the fee inside six to eight weeks, it's worth doing.

Most ten-person businesses I've done this exercise with land somewhere between 25 and 45 hours a week once the setup weeks are behind them. The recruitment agency above was the highest I've seen, and a small charity doing grant admin by hand was close behind it, mostly because grant applications are repetitive in structure even when the content changes each time, which is exactly the kind of case I cover on my page about AI for charities and grant writing admin.

What this means for a ten-person business owner reading this on a Tuesday night

If you run a ten-person business, you already know exactly which two tasks eat your team's week. You don't need a consultant to tell you that. What you need is someone who's done the setup weeks before, who knows which 60 hours are real and which are marketing fluff, and who can get you through the awkward first month faster than you'd manage alone. That's the entire value of hiring one, and it's worth being blunt about it rather than pretending the tool does the work by itself from day one.

The practical next step is on my hire an AI consultant page, and the engagement itself is described under how to find an AI consultant worth hiring.

If you want the guide for your own sector, see AI consultant by industry: what to automate first.

Frequently asked questions

How many hours a week can AI realistically save a ten-person business?

Most ten-person businesses save between 25 and 60 hours a week across the whole team once systems are set up, with the biggest wins coming from screening, follow-up, and appointment reminders rather than anything creative or judgement-based.

Is it worth hiring an AI consultant for a business with only ten people?

Yes, if your team spends more than a few hours a week each on repetitive admin like chasing, screening, or drafting, because the setup time saved by hiring someone experienced usually pays for itself inside six to eight weeks compared to figuring it out yourself.

How long before a ten-person business sees time savings from AI?

Expect the first three to six weeks to cost more time than they save while criteria are checked and outputs are trusted, with the real weekly time savings showing up consistently from around month two onwards.

What tasks save the most time for a small business team?

Booking and reminder tasks save the most, around 80 to 90 percent of the time previously spent, followed by follow-up chasing at 70 to 85 percent and first-pass screening at 60 to 75 percent, with document drafting and content creation saving less because a person still needs to check and edit the output.

Official documentation

• Anthropic's own documentation
• the NIST AI Risk Management Framework