akashuin

Blog

  • 71% of UK staff use unapproved AI. Only 32% worry about the data.

    71% of UK staff use unapproved AI. Only 32% worry about the data.

    Seven in ten UK employees have used an AI tool at work that their employer never approved. Only a third of them are worried about the client data they are feeding it. If you run a small agency or practice, assume that includes your team, and assume some of your clients’ information is already sitting in a free chatbot account you have never seen.

    That is not a scare line. It is what the numbers say when you put two recent studies side by side.

    What the research actually found

    Microsoft commissioned Censuswide to survey 2,003 UK workers in October 2025. 71% had used unapproved consumer AI tools at work, and 51% still do so every week. The uses are exactly what you would expect: 49% drafting emails and replies, 40% writing reports and presentations, and 22% on finance tasks. Only 32% said they were concerned about the privacy of the customer or company data going in.

    Now look at the other side of the desk. The government’s Cyber Security Breaches Survey 2025/26, published in April, found that 31% of UK businesses are using AI, adopting it or actively considering it. Of that group, just 24% have any cyber security practice in place to manage the risks from AI. The same survey found 46% of small businesses and 42% of micro businesses had a breach or attack in the last twelve months.

    Put those together and you get the real picture. Staff are using AI far more than their employers think, the employers who do know about it mostly have no rules, and the people typing are not worried. A separate survey of over 1,000 UK decision-makers this month found 88% believe unapproved AI is in use in their business and 51% fear staff are pasting sensitive data into it. Worth flagging that one came from an automation vendor with something to sell, but it lines up with everything else.

    Why your team does it

    Nobody on your team is being reckless for the fun of it. Microsoft asked why they use unapproved tools. 41% said it is what they already use in their personal life. 28% said their employer does not provide an approved alternative.

    That second number is the one I keep coming back to. In most of the small firms I work with, nobody has decided anything about AI. There is no policy, no paid account, no guidance. So a junior account manager with a deadline opens the ChatGPT tab they use for meal plans and pastes in the client brief. The work gets done faster, the client is happy, and the brief now lives in a consumer account that may be used to train future models, tied to a personal email address you have no control over.

    Shadow AI is not a discipline problem. It is what happens when a business leaves a gap and a deadline fills it.

    A practitioner writing on The Product Journey makes the point well: you cannot govern what you cannot see, and visibility has to come before any rule. Banning tools outright tends to push usage further underground, where it is harder to find.

    The specific risks for a small firm

    For an agency, accountancy practice or law firm, the exposure is concrete:

    • Client confidentiality. Most engagement letters and NDAs promise you will not share client information with third parties. A free AI account is a third party.
    • UK GDPR. If personal data goes into a tool with no data processing agreement, you are the controller and you cannot say where it went. That is a hard conversation with the ICO and a harder one with a client.
    • Leaving staff. When someone leaves, their personal AI account leaves with them, complete with months of chat history about your clients.
    • Bad output nobody checked. Unapproved use usually means unreviewed use. A hallucinated figure in a client report is your problem, not the chatbot’s.

    None of this means AI is the enemy. Microsoft’s own figures put the time saved at nearly eight hours a week per worker on admin. The goal is to keep that benefit while closing the gap.

    Try this: a one-week shadow AI fix

    You do not need an IT department for this. Here is what I would do in a firm of five to fifty people.

    1. Run a no-blame AI amnesty (Monday, 15 minutes). Send a short form: which AI tools do you use for work, what for, and what kind of information goes in? Say clearly that nobody is in trouble and the point is to get them better tools. You will be surprised by the answers, and that survey alone tells you where your real risk sits.

    2. Buy the business version of whatever they already use (Tuesday). If the answers say ChatGPT, get ChatGPT Team or Enterprise. If they say Copilot, check what your Microsoft 365 licence already includes, because many firms are paying for protected Copilot Chat and not using it. Business tiers typically exclude your data from model training and give you an admin console. This removes the 28% reason overnight: there is now an approved alternative, and it is the tool people already like.

    3. Write a one-page traffic-light rule (Wednesday). Not a twelve-page policy nobody reads. Three lines:

    • Green: public information, your own drafts, generic questions. Any approved tool.
    • Amber: client names, internal financials, project details. Company accounts only, never personal ones.
    • Red: personal data about individuals, anything under NDA, passwords, bank details. Not into any AI tool without sign-off.

    4. Close the personal accounts (Thursday). Ask everyone to move work chats into the company account and delete work content from personal ones. Add “AI accounts” to your leaver checklist so it happens every time someone goes.

    5. Check it in a month (put it in the diary now). Re-run the form. If people are still reaching for unapproved tools, find out what the approved one is missing and fix that, rather than adding another rule.

    If you would rather have someone do the discovery, test what your tools are actually doing with data and write the rules for you, that is what an AI security audit produces. But the steps above will get most small firms most of the way there in a week.

    The bottom line

    Three quarters of the UK businesses engaging with AI have no rules for it, and seven in ten workers are already using it anyway. The firms that get hurt here will not be the ones that used AI. They will be the ones that pretended their team was not using it. Pay for the tool, write the one page, and make the safe option the easy one. That is cheaper than any conversation you will have after a client’s data turns up somewhere it should not be.

  • UK firms use just 1.6 AI tools each. Here is how to pick the right two.

    UK firms use just 1.6 AI tools each. Here is how to pick the right two.

    UK businesses now average 1.6 AI tools each, up from 1.4 in 2023. That is the entire depth of three years of “AI transformation” for most firms: half a tool. Adoption tripled over the same period, from 12% to 35% of businesses with ten or more staff, according to the Office for National Statistics‘ Business Insights and Conditions Survey run in June 2026. Everyone signed up. Almost nobody went deep.

    I see this constantly with clients. A firm has ChatGPT on three people’s laptops, a Zapier-to-AI thing someone half-built in Q1, and a Notion AI toggle nobody remembers switching on. None of it talks to the others. None of it is anyone’s job to maintain. That’s not an AI strategy, it’s a pile of trial subscriptions, and 1.6 tools per adopter is what that pile looks like from the outside.

    The real problem isn’t too few tools. It’s no process for choosing them.

    Scott Pope at Nexthink put it well when the ONS numbers came out: adoption is “widening but not deepening”, and the cause is cultural as much as technical. I’d go further: most firms have never actually tested an AI tool against an alternative. They pick whatever a LinkedIn post recommended, run it for a fortnight, and either it sticks by accident or it quietly dies in a browser tab.

    Compare that with how a solo operator who actually depends on these tools does it. Angus Sewell wrote up his 2026 stack by running things head-to-head and keeping receipts. He didn’t swap his AI search provider because a blog told him to. He measured it: 87% retrieval accuracy against 73% for the built-in option, with the winner returning actual passages and dates instead of bare links. He dropped n8n entirely, not because automation platforms are bad, but because building a small custom AI skill turned out more adaptable for his actual workload than maintaining a flowchart of triggers. He even reversed a decision on custom frontends after clients told him plainly they preferred an off-the-shelf tool they wouldn’t have to think about.

    None of that is exotic. It’s just a testing habit most businesses skip because “trying an AI tool” gets treated as a five-minute favour to do between meetings rather than a decision worth ten minutes of comparison first.

    A framework that gets you past 1.6 tools without turning into a tool hoarder

    The goal isn’t to use more AI tools. Firms that bolt on six half-used subscriptions aren’t ahead of the 1.6 average, they’re just further into the pile. The goal is to run a real head-to-head before anything earns a permanent seat, and to be willing to cut what loses.

    1. Pick one task, not one tool, to start. Don’t ask “should we use Claude or ChatGPT”. Ask “who spends the most repeated hours per week on one narrow task”: drafting first-pass proposals, summarising call notes, chasing invoice status, writing the same three email types. That task is your test bench. A tool that’s brilliant in general but mediocre at your actual bottleneck task isn’t worth adopting yet.

    2. Run two candidates on the same ten examples. Take ten real pieces of past work, ten real emails you sent, ten real summaries you wrote, and get two tools to do the same job blind. Score them the boring way: how much did you have to rewrite before it was usable? Sewell’s Exa-versus-built-in-search comparison is the model here: a specific number, on a specific task, not a vibe.

    3. Give the winner thirty days with an owner, not a free-for-all. A tool that’s “available to the team” gets used by nobody in particular and dies quietly around month two. Name one person responsible for making it work, including chasing down the annoying edge cases where the output needs a human fix. That’s the difference between the tools that survive Sewell’s stack and the ones that don’t: someone kept forcing them to prove their worth.

    4. Cut the loser in writing. If a tool doesn’t win its head-to-head, cancel the trial and say so out loud in the team channel. This is the step firms skip most. Nobody wants to be the person who killed the AI initiative, so half-used tools survive by default rather than by merit, which is exactly how you end up as one of the 35% who “adopted AI” and one of the 1.6-tools-per-firm who never got anything real out of it.

    Do this with two tools properly rather than five tools badly, and you’ll be ahead of most of the 35% within a month. Adoption isn’t the scoreboard. A tool that’s actually load-bearing in someone’s daily work is.

    Where this tends to go next

    Once a firm has one AI tool genuinely embedded in a real workflow, not a demo, a habit, the next bottleneck almost never shows up as “we need another AI tool”. It shows up as “this works but someone still has to copy the output from one system into three others by hand”. That’s a process problem, not a tools problem, and it’s the point where a lot of the firms I work with hit a wall on their own. If that’s where you are, one AI habit that’s working, and a queue of manual steps still wrapped around it, that’s exactly what process automation is built to remove, and it’s worth thirty minutes to map out before you buy tool number three.

    Three years from now the ONS will publish another survey. My bet: the adoption percentage will look impressive again, and the average tools-per-adopter number will still be embarrassingly small, because most firms will still be choosing tools by vibe instead of by test. That gap is where the advantage actually sits: not in being first to try something new, but in being the firm that bothered to check whether it worked before keeping it.

  • AI can lift lead conversion 30%. 84% of small UK firms haven’t tried.

    AI can lift lead conversion 30%. 84% of small UK firms haven’t tried.

    AI lead scoring lifts conversion by roughly 30 percent, according to last year’s vendor benchmark data. Meanwhile only 16 percent of UK small businesses have deliberately deployed any AI technology at all. Those two numbers sitting next to each other tell you most of what you need to know about where the opportunity actually is.

    The adoption gap nobody quotes

    The headline figure everyone repeats is 29 percent: the share of UK businesses now using at least one AI technology, according to the Office for National Statistics’ Business Insights and Conditions Survey run in June 2026, up 8 points on the year before. It sounds like AI has gone mainstream. It has, but only for one size of company. Among businesses with 250 or more staff, adoption sits at 49 percent, up 13 points.

    Drop down to the businesses most agencies and service firms actually are, and the picture changes. Research from the Department for Science, Innovation and Technology puts deliberate AI deployment among UK firms with five or more employees at 16 percent, with a further 5 percent saying they have concrete plans to. That’s the real adoption rate for the businesses reading this post, and it’s roughly a third of the enterprise figure.

    The interesting part isn’t the gap in adoption. It’s the reason for it. When DSIT asked small businesses why they hadn’t deployed AI, cost came bottom at 23 percent and integration complexity next at 29 percent. Skills came second at 60 percent. The single biggest barrier, at 71 percent, was simply that they hadn’t identified a need for it.

    That’s not a tooling problem or a budget problem. It’s a use-case problem. Most small firms have never sat down and mapped a specific, repeatable job in their business onto a specific AI capability. Lead generation is one of the clearest places to do exactly that, because it’s already a repeatable process with a measurable outcome, which makes it easy to prove the case one way or the other.

    What an AI lead-gen system actually does

    “AI for lead generation” gets used loosely to mean everything from a chatbot widget to a fully autonomous pipeline. The version worth building is the latter: a system that runs the whole sequence without someone manually pushing it through each stage.

    A working pipeline typically covers five stages:

    • Discovery. Finding companies or contacts that match a defined signal (funding round, hiring pattern, tech stack change, a trigger event) rather than a static list.
    • Enrichment. Filling in the contact, company, and context detail that makes personalisation possible instead of generic.
    • Scoring. Ranking prospects so a human’s time goes to the ones actually worth it. This is the stage where that roughly 30 percent conversion lift cited in 2025 vendor benchmark data mostly comes from. Scoring doesn’t create leads, it stops good ones getting buried under bad ones.
    • Personalised outreach. A first draft that references something specific to the prospect, not a merge-tag template.
    • Follow-up. Re-engaging people who didn’t reply the first time, on a schedule, without someone remembering to do it.

    None of these stages need to be fully autonomous on day one. The mistake most small firms make isn’t picking the wrong tool, it’s trying to automate all five stages at once, stalling on the integration work, and concluding “AI doesn’t really apply to us.” That’s exactly the 71 percent barrier showing up in practice, not just in a survey.

    This is also the point where hiring versus automating stops being a philosophical question. A pipeline like this is genuinely multi-step work: research, judgement calls, drafting, timing, which is what makes it a good candidate for an autonomous agent rather than a single tool bolted onto a spreadsheet. The difference between a chatbot and an agent here is whether the system can carry a prospect from discovery through to a drafted, personalised message without a person moving it along at every step.

    Try this: pick one stage, not five

    If you run a small firm and want to actually test whether AI helps your pipeline rather than just reading about it, do this over one week:

    1. Pick your slowest stage, not your favourite one. Time how long each of the five stages above currently takes, from someone opening a list to a message actually going out. Automate the one taking longest first, usually enrichment or follow-up, rarely discovery.
    2. Measure reply rate, not volume. More leads processed faster means nothing if reply rates fall because personalisation got thinner. Track reply rate for the two weeks before you change anything and the two weeks after. If it drops, the automation made the messages worse, not just faster; fix that before scaling it.

    Both of those are things you can do this week with what you already have, before spending anything on new tooling.

    The honest version

    The case for AI in lead generation isn’t that it will replace the judgement calls. Deciding which prospect is actually worth a call still needs a person who understands the business. It’s that scoring, enrichment, and follow-up are exactly the parts of the job that don’t need judgement, and are also the parts most small teams do worst, because they’re tedious and easy to let slip. Automate those three and the person doing outreach gets to spend their time on the 20 percent of prospects that were always going to close, instead of working evenly through a list where most of it never had a chance.

    The 29 percent adoption headline was never really about firms like most of akashuin’s readers. The 16 percent figure is. If you’re going to be one of the businesses that moves from that 16 percent to something higher, lead generation is one of the few places where you can prove the return in weeks, not quarters. If you’d rather have that pipeline built for you than build it stage by stage, that’s what an autonomous agents engagement is for.

  • Everyone’s writing with AI. Platforms just started punishing it.

    Everyone’s writing with AI. Platforms just started punishing it.

    Substack told its writers this year that it will start demoting and removing “AI slop” from its recommendation engine. No warning shot, no grace period. If a post reads like generic filler with no point of view behind it, the algorithm now buries it, even when a human technically wrote it. That’s the moment content marketing changed, and most agencies haven’t clocked it yet.

    The gap nobody’s talking about

    Marketing teams have gone all-in on AI for content. HubSpot’s 2026 State of Marketing report puts AI use at 91% of teams, up from 63% the year before, which is fast adoption for anything in this industry. Idea generation, first drafts, multi-asset campaigns, product copy: all largely AI-assisted now.

    Here’s the part that should worry you more than the adoption number. Only 41% of marketers say they can confidently prove that AI is delivering ROI, down from 49% the year before. Adoption went up. Confidence in the results went down. That’s not a rounding error, it’s a trend line pointing the wrong way, and it’s exactly the pattern you’d expect if a lot of teams adopted the tool without adopting the discipline that makes it pay off.

    The UK numbers tell the same “everyone’s doing it, few are doing it well” story. The ONS’s June 2026 survey of AI in UK businesses found 35% of businesses with ten or more staff now use at least one AI technology, up from around 12% in late 2023. Within that, 18% are using large language models specifically to generate text, the single most common use case. But adoption is shallow: the average number of AI tools an adopting business actually uses has crept from 1.4 to just 1.6 over the same period. Widespread use. Thin evidence it’s actually working, and barely any sign firms are going deeper once they start.

    Why the ROI number is falling, not rising

    Two things are happening at once. First, governance hasn’t kept pace with output speed. Concerns about legal, compliance and brand review grew 3.4 times year over year in the same report, the fastest-growing worry in the whole survey. AI can produce a week’s worth of content in an afternoon. Most review processes were built for a much slower publishing cadence, so a lot of that output goes out unchecked simply because nobody had time to look at all of it.

    Second, and more basic: a lot of what’s going out is genuinely bad. Not wrong, just forgettable. It reads like the median of everything ever written, smoothed into a paste and published at volume. Readers can feel it even when they can’t say why, and platforms are starting to detect it algorithmically too. Substack’s policy shift is specific about the target: generic listicles, recycled takes, posts with no discernible point of view. Using AI for research or editing is fine. Publishing its unedited median opinion is what gets punished, and that distinction is the one most agencies still haven’t built into their workflow.

    Adoption solved the speed problem. It didn’t solve the taste problem, and taste is the part a platform can now detect.

    What “doing it well” actually looks like

    The firms in that 41% who can prove ROI aren’t using less AI than everyone else. They’re using it with a visible layer of human judgement sitting on top of the output, applied at two specific points: before a draft gets written, and before it goes live. Neither point requires slowing production down much. Both require someone to actually own the decision, rather than letting a tool’s default output become the published version by default.

    Ruben Hassid, whose “How to AI” newsletter has become one of the more widely read practical AI guides for non-technical readers, has a method worth stealing for the first point. Before you generate anything, build a running document (he calls it an anti-ai-writing-style file) that lists every word and phrase you’ve come to recognise as an AI tell: “unlock,” “let’s dive in,” the “it’s not X, it’s Y” sentence pattern, overused dashes, whatever your own ear catches. Feed that file into your prompt every time. It’s a five-minute setup that pays off on every piece you generate afterwards, because it forces the model to write in a voice instead of the average voice.

    Try this

    Two changes you can make this week, no new tools required:

    1. Build the ban list. Open a doc, name it whatever you like, and every time an AI draft uses a word or construction that makes it read as machine-written, add it. Paste the list into your system prompt or custom instructions before every generation. It gets more useful every week you keep it, and it’s the single cheapest fix for the “slop” problem, because most AI tells repeat across dozens of pieces once you start noticing them.

    2. Add one gate before publish, not before drafting. Governance concerns are spiking because review hasn’t scaled with output, so don’t try to review everything equally hard. Pick the one question that actually catches bad content: “could a named person on our team defend every claim in this piece if a client asked them to?” If the honest answer is no, it doesn’t go out under your brand, however fast it was to produce. Assign that gate to one person by name, not to “the team”, or it quietly stops happening within a month.

    The opinion part

    AI content tooling is not the differentiator anymore. At 91% adoption it was never going to stay one for long. The differentiator is whether a firm has built the fifteen minutes of human judgement around each piece that keeps it out of the “slop” bucket Substack is now actively burying. That’s a process problem, not a technology problem, and it shows up first in a team’s skills and habits rather than in its toolkit. If you’re not sure whether your team’s AI habits would survive that kind of audit, that’s what an AI skills and adoption review is for: not to add more tools, but to find out where the judgement layer is thin before a platform, or a client, finds it for you.

  • The AI that pays off isn’t a chatbot. It’s a rule.

    The AI that pays off isn’t a chatbot. It’s a rule.

    Most UK service firms that say they have “done AI” have installed a better search box. They pay for ChatGPT or Claude, a few people use it to draft emails and summarise documents, and that is where it stops. The Office for National Statistics has the numbers: around 35% of UK businesses with ten or more staff now use AI, but the average adopter runs just 1.6 different AI technologies, barely up from 1.4 in late 2023. Adoption is wide and shallow. The return on investment is somewhere else entirely.

    What firms actually bought

    The ONS breakdown for June 2026 is revealing. Large language models sit at 18% of businesses, visual content creation at 16%, machine-learning data processing at 12%, robotics at 2%. “Improving business operations” is the most common stated purpose, at roughly 60% of AI users. Yet only 11% of firms have given more than half their workforce any AI training, and 41% say they face no barriers to adoption at all.

    Read those last two together. The tool is on the desk, almost nobody was taught to build anything with it, and most firms do not believe anything is stopping them. They simply have not done the work. And the work being skipped is process automation: not a chatbot that answers a question, but a system that takes a recurring, multi-step job and runs it from start to finish, with a person checking the part that carries risk.

    Where the return actually is

    People who build these systems for a living are blunt about it. The money is in deterministic, rule-based workflows, and you do not need a model for most of it. One automation practitioner writing on Substack lays out three tiers, and the order matters:

    • Deterministic workflows. Predictable, rule-based, no model involved. A form submission creates the project, sets up the folder, sends the kickoff email and books the calendar invite.
    • AI-enhanced workflows. The model handles one small judgement inside an otherwise fixed process. It reads an enquiry and picks which template applies, or turns messy call notes into a standard summary. This covers roughly half of real tasks.
    • Autonomous agents. Genuinely open-ended, and genuinely unpredictable. Only worth attempting once the first two tiers are solid.

    A concrete version. A recruitment firm signs a new client, and someone spends an hour and a half creating the folder structure, copying the contract details into the CRM, drafting the welcome email, setting up the shared tracker and scheduling the intake call. Every field in that sequence has a fixed source. None of it needs a model. Built as a plain workflow triggered by the signed contract, the same sequence takes about four minutes of review, and the person who used to do it gets that ninety minutes back every time.

    Rule-based automation of this kind returns 30% to 200% in the first year, according to that same analysis. The reason it works is unglamorous. It removes hours of copy-paste admin that a person was doing every week, and those hours have a known cost.

    A chatbot answers a question. A rule finishes a job.

    Why the boring version gets skipped

    Autonomous agents are the interesting part, so that is where people start. They hit the unpredictability, watch it break on something a rule would have handled, and conclude automation is not ready yet. As that practitioner puts it: “They start at the top, hit the unpredictability, and conclude automation doesn’t work.”

    Meanwhile the client, or your own team, never wanted an agent. They wanted the client onboarding sequence to stop eating ninety minutes every time someone signs. They wanted the monthly report to assemble itself. Ruben Hassid, who writes the How to AI newsletter for a large non-technical audience, frames the same idea as spotting the repeated pattern, naming the context it needs, and turning it into a reusable checklist. Start there, not with the moonshot.

    Try this

    Four steps, in order, that you can start this week.

    • Run the two-of-four test. Take a task and check it against four criteria: repetitive, time-consuming, error-prone, scalable. If it meets at least two, it is a candidate for automation. If it meets none, leave it alone. This stops you automating things that were never the problem.
    • Count the hours before you build anything. Time the task honestly across a normal week, then multiply by the hourly cost of whoever does it. That single number is both your business case and your measure of success. The goal is “fifteen hours recovered this month”, not “we deployed AI”.
    • Build rule-first. Map every step. Automate the ones with a fixed answer using plain tooling: Zapier, Make, a short Google Apps Script, or the workflow builder already inside your CRM. Only bring a model in where a step genuinely needs judgement, such as classifying an inbound message or compressing notes into a set format.
    • Keep a person on the risky step. Anything that sends an external email, moves money, or changes a client record should pause for human approval before it fires. That approval gate is what makes the whole thing safe to switch on and leave running.

    Working out which of your recurring workflows are worth automating, and in what order, is most of the effort. It is also exactly what our process automation service produces if you would rather not run the exercise in-house: a shortlist of processes ranked by hours saved, then the builds that clear them.

    The point

    If you run a UK service firm and you have “done AI” this year, ask one question. Have you automated a single process from end to end, or have you just handed people a smarter search box? The first compounds quietly every month. The second is a subscription. The firms pulling ahead are not the ones with the cleverest agents. They are the ones who found the ninety-minute job that happens every week and turned it into a five-minute job.

    Sources: ONS, Artificial intelligence in UK businesses, 2023 to 2026; “You’re building agents. They wanted a rule.”, Substack; How to AI by Ruben Hassid.

  • Everyone added AI to cold email. Reply rates barely moved.

    Everyone added AI to cold email. Reply rates barely moved.

    Everyone bolted AI onto their cold email this year. Reply rates, on average, did not move.

    Instantly’s 2026 Cold Email Benchmark Report, built from billions of interactions across thousands of active workspaces over 2025, puts the overall average reply rate at 3.43%. That number is barely different from where cold email sat five years ago, before every outbound tool shipped an AI writer. Meanwhile the British Chambers of Commerce’s March 2026 research shows 54% of UK firms now actively using AI, up from 35% in 2025 and 25% in 2024. Adoption tripled in two years. The metric that actually pays the bills stayed flat.

    That gap is the real story, and it’s a useful one if you run outreach for an agency or a small service firm. AI didn’t fail at lead generation. Most people used it to do the same thing faster, not to do a different thing.

    Where the reply rate actually moved

    The BCC data has a second number worth sitting with: of the 54% of firms using AI, only 11% use it extensively to automate or streamline operations. The rest are dabbling: a chatbot here, an AI-drafted email there, without changing the underlying process. Same list, same generic template, same “Hi {firstName}, hope you’re well” opener, just typed by a model instead of a person.

    Instantly’s own benchmark shows what the other tier looks like. Their top 10% of senders hit 10.7%+ reply rates, three times the average. What separates them isn’t a better AI model. It’s what the AI is pointed at. The consistent thread across the top performers, and across separate research from Sendr, Unify GTM and Martal’s 2026 cold email data, is signal-based targeting: emailing people because something specific just happened (a hire, a funding round, a tool switch, a website visit) rather than because they match a firmographic filter.

    The mechanism is simple. A generic list gets a generic message, and the prospect can tell in one line. A signal-triggered message references something true and current about their business, which is the entire definition of relevance. AI is very good at drafting that message once you hand it the signal. It is not good at inventing the signal for you, and most tools people bought this year only automate the drafting.

    Adding AI to a bad list gets you a faster bad list.

    The maths behind fewer, sharper emails

    Volume was the old lever: send more, book more. Signal-based sending inverts that. A team sending 200 emails triggered by a real signal at a 20% reply rate books roughly the same number of conversations as a team sending 1,000 generic emails at 3%, using a fifth of the volume, a fraction of the domain-warmup risk, and far less time spent by whoever’s writing follow-ups. Fewer, better-targeted emails compound: deliverability holds up longer because you’re not burning domains on blast volume, and every reply is a genuinely interested prospect rather than someone confused about why they got the email.

    This is also why “AI for lead gen” as a category gets a mixed reputation. Bought as a volume multiplier, it multiplies whatever was already mediocre about the list and the message. Bought as a targeting and drafting layer on top of real signals, it does what people hoped AI would do for outreach in the first place.

    Try this

    Two changes you can make this week without buying anything new:

    1. Benchmark your own reply rate against 3.43% before you touch AI copy. Pull your last 90 days of outbound. If you’re below 3.43%, the problem is almost never the writing; it’s the list. No amount of AI polish fixes a list built on industry and headcount filters alone. Fix the list first.

    2. Pick three trigger events and test a small signal-based batch against your usual send. New hire in a relevant role, a funding announcement, a tool or vendor change you can detect (job posts mentioning a competitor’s product are an easy one). Send 50–100 emails against real signals within 48 hours of the trigger (timeliness matters roughly as much as the signal itself, since relevance decays fast) and run it against your normal batch as a side-by-side. Most firms that do this find the signal batch outperforms 3–5x, which matches what the wider research above shows. You don’t need a fancy intent-data platform to start: a saved LinkedIn search and a Google Alert cover the first three signals for free.

    Once that’s proven at small scale, the actual bottleneck becomes catching the signals reliably and getting the drafted email into a rep’s outbox within the 48-hour window — which is a process problem, not a copywriting one. That’s the part worth automating properly rather than doing by hand every morning: it’s exactly the kind of repetitive, time-sensitive workflow we build for clients at Ishigai’s process automation service — watching for the signal, drafting against it, and queuing it for a human to approve, so nothing sits stale past the window that actually matters.

    The opinion part

    The 3.43% average isn’t a ceiling on what AI can do for outreach. It’s a snapshot of what happens when a fast tool gets pointed at a slow, generic process and everyone assumes the tool did the work. It didn’t. The list, the signal and the timing did the work; the AI just typed it up quicker. Firms chasing the 10.7% tier aren’t using smarter AI. They’re using AI on a sharper input. Fix what goes in before you worry about what writes it.

  • AI budgets are up 88%. ROI isn’t keeping pace.

    AI budgets are up 88%. ROI isn’t keeping pace.

    Nine in ten UK businesses raised their AI budgets this year. Fewer than a third can point to a return on it. That gap between rising spend and flat results is the real AI story for UK agencies right now, and it isn’t a technology problem.

    According to Spicy Advisory’s 2026 UK SMB AI adoption guide, 85-91% of UK organisations increased AI spend this year, yet only 31% report positive ROI. Seventy-seven percent saw no measurable revenue change at all. Where AI does pay off, it shows up as time saved rather than top-line growth, which would be fine, except most firms aren’t measuring time saved either, so the return is invisible even when it’s real.

    ONS figures from June 2026 point at why. The average AI-adopting UK business uses just 1.6 AI tools, and only 10% report extensive use of any of them. Adoption has tripled since 2023, but depth hasn’t moved much at all. Only 15% of businesses in the 0-9 employee bracket — where most agencies sit — use AI at all, against 68% of large firms. UK firms aren’t short of AI. They’re short of depth.

    There’s a skills story sitting underneath both numbers. Over 60% of firms name the skills gap as their primary barrier, and only 11% have trained more than half their workforce in AI skills. Buying a licence isn’t the same as building the muscle to use it well, and most firms are stopping at the purchase.

    More tools, same problem

    Tool fragmentation is Spicy Advisory’s second-named barrier to AI ROI, right behind the skills gap. It’s a pattern I run into constantly with clients: a Copilot licence bought for the whole team, a ChatGPT subscription three people actually open, a Claude account someone on the leadership team set up after a conference, and none of them wired into an actual workflow. Everyone in the building is “using AI.” Nobody’s using it for anything specific enough to measure.

    Success in 2026 isn’t about using every AI tool. It’s about choosing the right tool for the right task and going deep on it, as the Stop Chasing Every New Tool newsletter put it this year.

    That’s the uncomfortable part for a lot of firms: the fix isn’t procurement, it’s subtraction. Before evaluating a fourth or fifth tool, work out what the first three are actually doing for you, if anything.

    Try this: a 20-minute tool audit

    You don’t need a consultant for the first pass. Block 20 minutes and do this:

    1. List every AI tool with an active licence. Finance can usually pull this off the card statement faster than IT can list it from memory.
    2. Write down the one workflow each tool touches — not “content” or “admin”, but the actual task: “drafting first-pass client proposal decks”, “summarising onboarding calls”.
    3. Cut or pause anything without a named workflow attached. If nobody in the room can say what a tool is for, it isn’t producing ROI, measured or not, and it’s a licence you can reclaim this afternoon.
    4. Pick the tool touching your highest-volume repetitive task and go deep on it. Build the prompt library, the template, the checklist, whatever makes the fiftieth use faster than the first. That’s where the actual time saving lives, and it compounds; a tool used once a week never gets good enough to notice.

    Most firms find step 3 uncomfortable, because “we’re paying for it, so we must be using it” feels true even when the usage logs say otherwise. Check the logs. Most seat-based AI subscriptions show login frequency somewhere in the admin panel, and it’s usually a harder number than anyone expects.

    Then close the loop: whichever tool you keep going deep on, write down what you’re measuring before you start: hours saved on that one workflow, turnaround time, or output volume. Not “productivity”, a specific number tied to the specific task from step 2. That’s the difference between joining the 31% who can show ROI and staying in the 69% who are pretty sure it’s helping but can’t prove it.

    The standard stack, plus one

    Spicy Advisory’s guidance for UK SMBs is blunt and, in my experience, correct: standardise on Microsoft 365 + Copilot or Google Workspace + Gemini for the baseline, add one external assistant — Claude or ChatGPT — for the work your core suite doesn’t cover well, and stop there until you can show the first two are earning their keep.

    That’s a smaller stack than most firms are currently running, and that’s the point. Depth over breadth isn’t a slogan here: it’s the difference between the 31% seeing ROI and the 69% who aren’t. The 31% aren’t using more tools. They’re using fewer tools for longer, on narrower, higher-volume tasks, until the tool actually knows the job.

    If the audit above turns up five half-used subscriptions and nobody in the business owns the decision of what to cut or what to go deep on next, that’s exactly the kind of prioritisation call worth getting an outside, structured view on. That’s the whole point of an AI opportunity audit: not another tool recommendation, but a ranked list of where AI actually pays off in your specific workflows, and where it’s just spend.

    The opinion part

    AI budgets will keep rising into 2027 whether or not ROI follows: the EY and BCC data both point that way, and nobody wants to be the firm that “isn’t doing AI.” But the 69% gap between spend and return isn’t going to close by adding tool number four. It closes by someone in the business deciding which one tool gets used until it’s actually good, and which three get switched off. That decision is boring, it takes 20 minutes, and almost nobody is making it.

  • UK firms tripled their AI use. Few changed how they work.

    UK firms tripled their AI use. Few changed how they work.

    The Office for National Statistics published figures this month that should give pause to anyone selling AI to British businesses, and anyone buying it. AI use among UK firms with ten or more employees has roughly tripled since late 2023, climbing from about 12% to 35%. On the face of it, that reads like a technology winning.

    Then you look at the depth. The average number of AI tools per adopting firm moved from 1.4 to 1.6. Only 10% of adopters call their own usage “extensive”. Just 15% say more than half their staff use AI daily. And only 11% report that more than half their workforce has had any AI training at all.

    Three years of enthusiasm. Two-tenths of a tool.

    Adoption is not the same as change

    The gap gets wider the closer you look. The British Chambers of Commerce, working with Atos, put UK SME AI adoption at 54%. The government’s own DSIT adoption research finds roughly 16% of firms have made a strategic AI deployment. Both numbers can be true, because they are measuring different things. One counts firms where somebody, somewhere, has opened ChatGPT. The other counts firms that have actually rebuilt a process around it.

    The ONS data makes the same point in a different way. Nearly 60% of adopting firms use AI to make existing operations more efficient. Fewer than one in five use it to build new products, reach new markets, or do something they could not do before. Around half report no change in headcount, which is a polite way of saying nothing structural happened.

    Most firms have bought a subscription. Very few have changed a workflow.

    I see this constantly. A recruitment agency tells me they are “using AI” and what they mean is two consultants paste job descriptions into a chatbot and tidy the output. That is not nothing. It saves them twenty minutes a day each. But it is not the thing that changes the shape of the business, and it will not survive either of them leaving.

    The barrier is not what people assume

    Here is the finding I keep coming back to. When the ONS asked firms what was holding them back, 41% said nothing at all. No regulatory problem, no budget problem, no skills crisis. Nothing.

    Among firms that did name a barrier, the top answer was lack of expertise, cited by up to 18% of businesses with 100 to 249 staff.

    Put those together and you get an uncomfortable picture. A large share of British firms are not blocked. They simply do not know what to do next. They have the budget, the permission and the appetite, and they are stuck at the point of asking which problem is worth solving. That is a very different constraint from the one most AI vendors are selling against.

    It also explains the tool count. If you do not know where the value is, buying one more tool feels like progress. It is measurable, it is quick, and it produces something you can mention in a board meeting. Redesigning how quotes get produced does not.

    What the productivity numbers actually promise

    The upside is real, which is what makes the shallowness frustrating. Public First’s analysis for Google estimates AI tools could lift UK SME productivity by around 20%, roughly an extra working day a week, worth up to £198 billion across the economy. A 2025 government study found 56% of businesses using AI reported higher employee productivity.

    Notice the framing though. An extra day a week is not what you get from a chatbot subscription. It is what you get when a process that took four hours takes forty minutes, permanently, whether or not the person who set it up is in the office.

    Try this: the two-week audit you can run yourself

    You do not need a consultant to find your first real candidate. You need a week of honest observation and a spreadsheet. Here is the version I use before quoting anyone.

    1. Log the repeats, not the annoyances. For five working days, have your team note any task they did more than three times that week that followed roughly the same steps each time. Not what irritated them, what repeated. The two lists overlap less than you would think, and the repeats are where automation pays.

    2. Score each one on three columns. Hours per month. How much judgement it needs, high, medium or low. And whether the inputs live somewhere a computer can reach, an inbox or a CRM counts, someone’s memory does not. Anything with high hours, low judgement and reachable inputs is your shortlist. That is usually three or four things, and they are rarely the ones people guessed.

    3. Kill your favourite idea first. Whatever you were excited about before you started, check it against the three columns. If it needs high judgement or the data lives in someone’s head, it is a bad first project regardless of how good it sounds. Starting with the hard one is the most common way these efforts die.

    Do that honestly and you will end up with a ranked list of two or three processes with hours attached. That list is worth more than most AI strategy documents, because it is about your business rather than the category. If you would rather not run it yourself, that is more or less what an AI opportunity audit produces, but the method is not a secret and you are welcome to it.

    The uncomfortable conclusion

    The ONS numbers will be read two ways over the next few months. Vendors will use the 35% to argue that everyone is doing this and you are behind. Sceptics will use the 1.6 tools and the flat headcount to argue it is all hype.

    Both readings miss it. The interesting group is the 41% who report no barriers at all and still have not moved. They are not waiting for the technology to improve or the regulation to settle. They are waiting for someone to tell them which of their problems is worth pointing this at.

    My honest opinion, having built these systems for agencies and service firms: the firms that pull ahead over the next two years will not be the ones with the most AI tools. They will be the ones who picked two processes, rebuilt them properly, and left the rest alone. Breadth is what the statistics measure. Depth is what pays.