68% of UK Businesses Now Use AI, Here Is What They Automated First

In 2022, just 34% of UK businesses used any form of AI. Today, that number is 68%.

That's a historic acceleration. But here's the catch: only 16% of those businesses use AI strategically. The rest are experimenting with chatbots, copying competitors, or chasing whatever went viral on LinkedIn last week.

I've deployed automation for 120+ businesses across 15+ industries. The ones that win don't start with what's shiny. They start with what bleeds time.

What UK Businesses Are Actually Automating (Ranked by First Deployment)

I analysed the past three years of deployments across my client base. These are ranked by how often they appear as the first automation a business chooses to build:

1. Customer Service (Email & Chat Triage), 34% of first automations

A customer emails. An AI reads it. If it's a refund request, it routes to finance. If it's a feature question, it routes to support. If it's feedback, it goes to the product team.

Before: A human opens every email. Reads it. Routes it manually. The busy season costs them 20+ lost emails and a frustrated team.

After: 89% of emails route themselves. Your team handles the 11% that actually need human judgment.

The maths: A 15-person B2B SaaS company gets 400 emails/day. Customer service triage takes about 2.5 hours. Automate it, and you save 12.5 hours/week. At £30/hour, that's £375/week. The automation template costs £97/month. It's positive ROI by week one.

This is why customer service is first: the win is immediate, measurable, and you can see it the next Monday morning.

2. Data Entry & Manual Reporting, 24% of first automations

A sales team uses Pipedrive. Finance uses Xero. Nobody's talking to each other. Every month, someone manually exports from Pipedrive, cross-checks against Xero, and builds a reconciliation report in Excel.

That job takes 6-8 hours. It happens monthly. It's boring. It's error-prone. And it's exactly the kind of work AI should kill.

Deployment: A workflow reads new deals from Pipedrive, checks them against Xero, flags discrepancies, and builds a dashboard. No human except to review exceptions.

Result: The task that took 8 hours now takes 30 minutes. One person gets 7.5 hours back. Every month. Forever.

Data entry variations I've seen automated:

• Invoice-to-accounting entry (PDF to database)
• Expense claims to finance (Slack to Xero)
• Survey responses to CRM (Typeform to HubSpot)
• Customer data deduplication (clean up database garbage)

The pattern: If you're copy-pasting from one tool into another, that's an automation waiting to happen. The ROI is boring but it's real.

3. Lead Qualification, 19% of first automations

Your sales team books 40 discovery calls/week. 16 of them are people who don't fit your product. They have no budget. Or the wrong use case. Or they're just kicking tyres.

Those 16 hours are gone. No deal. No learning. Just gone.

An AI qualification bot answers the intake form. It asks three questions: What's your annual revenue? What's the problem you're solving? How soon do you need to fix it? Then it routes warm leads to sales and politely declines the rest.

Result: Sales now book 32 calls/week instead of 40. But 28 of them are qualified. The booking rate drops by 20%. The conversion rate jumps by 60%.

One recruiting firm I worked with automated candidate screening. Their team reviewed 3x more applications in half the time. They filled roles 3 weeks faster.

4. Compliance & Quality Checks, 14% of first automations

A healthcare provider needs every patient note reviewed for completeness before it gets filed. A legal firm needs contracts checked against compliance templates. An agency needs client deliverables signed off before sending.

These aren't creative tasks. They're checkboxes. And they're slow.

An AI audit runs through the document, checks it against a checklist, flags gaps, and routes it back to the human for 10 seconds of review instead of 5 minutes of creation.

Time saved: 40-60% on the review cycle. Zero risk because a human still signs off.

5. Content & Copy Generation, 9% of first automations

This is where most teams want to start. But it's not where they should. It's the last thing, not the first.

Why? Because every hour saved on customer service or lead qualification compounds every day. Every hour saved on copywriting maybe saves something next month.

The businesses that automate copy first often waste months perfecting AI-generated blog posts when they could have automated their customer intake and freed up the person writing those blog posts.

The Pattern: Automate the Painful, Repetitive, Measurable Work First

The common thread across all successful first automations is this: they are boring, specific, and the time saved is immediately provable.

You can measure customer service triage by emails processed. You can measure data entry by hours saved and errors eliminated. You can measure lead qualification by booking rate and conversion rate.

You cannot easily measure "better copy," which is why every struggling AI project starts there.

The second pattern: the successful teams define success before they build. They say "customer service triage saves 8 hours/week," and they measure it. The struggling teams deploy something and hope they saved time.

The third pattern: they pick something that's broken right now. Not something that could be better in six months. Something that's bleeding time this week.

Why Most Teams Pick Wrong (And How to Avoid It)

The failures I see fall into three categories:

1. The Excitement Trap. A team watches a demo of AI writing blog posts and thinks "that's amazing, let's start there." It feels modern. It feels like AI. It's not the biggest pain point. But it feels like the future. Six months later, they've generated 50 blog posts and zero leads, and their customer service team is still manually triaging 400 emails/week.

2. The Vagueness Trap. "Let's automate customer service." Customer service is massive. It spans email, chat, phone, social, support tickets. Where do you start? The teams that win say, "We're automating email triage first. Success is: 80% of emails route themselves within 2 weeks." The teams that lose say, "We're automating customer service" and then wonder why they're still not sure what they're building.

3. The Complexity Trap. A team decides to automate their entire customer journey at once. Intake, qualification, onboarding, support, offboarding. That's not an automation. That's a product rebuild. It fails because it's too big, takes too long, and nobody agrees on what success looks like. The teams that win automate customer service intake first. If it works, they add qualification next. One step at a time.

How to Choose Your First Automation (The Diagnostic)

Ask your team these five questions. Whichever has the highest score is your first automation.

1. How many hours/week do we spend on this? If it's less than 4 hours, skip it for now. If it's 8+, it's a candidate.

2. How repetitive is it? If you're doing the exact same task 20+ times/week, automate it. If it's different every time, wait.

3. How measurable is the success? If you can measure time saved or errors prevented, deploy it first. If success is fuzzy, delay it.

4. How broken is it right now? Is it causing delays? Lost deals? Customer complaints? If yes, go first. If it's working fine but could be better, wait.

5. How much will the team believe in AI after we deploy this? Pick something where the win is undeniable. If you automate email triage and save 10 hours/week, the entire company notices. If you automate copy and save 2 hours/week, only the writer notices.

Score each from 1-5. The area with the highest combined score is your target.

FAQ

Don't we need to start with strategy before choosing what to automate?

You do need strategy, but not before you pick. You need it during. Define what success looks like before you build: time saved, errors prevented, quality improved. If you can't measure it, you don't yet know what you're building. But you don't need a 12-month roadmap. You need a two-week definition of done.

What if we're not sure if something can be automated?

If it's repetitive and humans are doing it, it can be automated. The question isn't whether it can be done. It's whether it's worth doing. That's where the diagnostic above helps. Some things can be automated but shouldn't be (yet). Start with what's painful first.

Can we automate more than one thing at once?

Only if you have the team to support it. Most teams have one person leading automation. They can champion one well or try to lead three and fail at all of them. Deploy one, measure it, celebrate it, then deploy the next. The second automation is always easier than the first.

What if our first automation doesn't work?

It usually will. The failure rate is less than 5% for the automations in this list if you measure success upfront. The failures I see are always teams that built something without defining what "done" looks like. If you skip that step, of course you won't know if it worked.

How long does the first automation take to build?

Email triage: 2-3 weeks. Lead qualification: 3-4 weeks. Data entry: 2-3 weeks. Compliance checking: 2-4 weeks. If a project goes beyond 6 weeks, something went wrong. You either chose something too complex or you're trying to build version 2.0 instead of version 1.0. Start small. Win fast. Expand later.

Richard Batt has delivered 120+ AI and automation projects across 15+ industries. He helps businesses deploy AI that actually works, with battle-tested tools, templates, and implementation roadmaps. Featured in InfoWorld and WSJ.

Frequently Asked Questions

How do I know if my business is ready for AI?

You are ready if you have at least one process that is repetitive, rule-based, and takes meaningful time each week. You do not need perfect data or a technical team. The AI Readiness Audit identifies exactly where to start based on your current operations, data, and team capabilities.

Where should a business start with AI implementation?

Start with a process audit. Identify tasks that are high-volume, rule-based, and time-consuming. The best first automation is one that saves measurable time within 30 days. Across 120+ projects, the highest-ROI starting points are usually customer onboarding, invoice processing, and report generation.

How do I calculate ROI on an AI investment?

Measure the hours spent on the process before automation, multiply by fully loaded hourly cost, then subtract the tool cost. Most small business automations cost £50-500/month and save 5-20 hours per week. That typically means 300-1000% ROI in year one.

Put This Into Practice

I use versions of these approaches with my clients every week. The full templates, prompts, and implementation guides, covering the edge cases and variations you will hit in practice, are available inside the AI Ops Vault. It is your AI department for £97/month.

Want a personalised implementation plan first? Book your AI Roadmap session and I will map the fastest path from where you are now to working AI automation.

Already know what you want to automate? Inside the Vault you'll find customer service triage templates, lead qualification workflows, data entry automations, and compliance checking prompts. Pick one, deploy it, measure it, move on.