The Best Companies Are One Quarter Into AI. You’re Not Behind Yet

What One of the Best-Run Companies Just Told Us

If you have spent 2026 feeling like every other business has quietly become an AI powerhouse while yours is still arguing about which tool to license, the announcement Acquisition.com posted on LinkedIn in July should recalibrate you. This is a company famous for operational rigour, with more resources, better systems and stronger execution discipline than almost any SME will ever have. And by its own account, it began its serious, structured, company-wide AI push in April 2026, one quarter ago.

Not 2023. Not “we've been doing this for years.” April. They surveyed their team's AI proficiency at the start of the quarter, ran a deliberate programme for 90 days, surveyed again at the end, and published the before-and-after. The programme itself was almost boring in its simplicity: structured training sessions, weekly office hours where anyone could bring a problem, a peer champion in every department, and a stated commitment that the whole company was in: no exemptions for seniority, no opt-outs for the sceptics.

That last point matters more than it looks. Most corporate AI initiatives quietly excuse the leadership team and anyone who claims to be too busy. Acquisition.com framed theirs around the principle of “everyone in, no one left behind” (their words), and the results suggest the framing did real work.

The Numbers: 90 Days, Measured

Here is the before-and-after they published, as self-assessed proficiency across six skill areas, April to July 2026. All figures are from Acquisition.com's public announcement.

Two readings of this table are worth your time. The obvious one is the size of the gains: 30- and 40-point jumps in a single quarter. The less obvious one is the starting points. In April 2026, a team at one of the world's most admired operating companies rated itself 18% on agentic AI and 29% on workflow redesign. Those are the skills that actually change how a business runs, and even they were starting near zero. Whatever your team's honest baseline is right now, it is probably not far off where theirs was.

And notice where the biggest gain landed: workflow redesign and automation, up 41 points. The hardest, most valuable skill moved the most. That is what tells you this is a training-and-reps problem, not a talent problem. Nobody hired their way to that number in 90 days. They practised their way to it.

Why It Worked: Five Mechanisms

Strip the announcement back and there are five mechanisms doing the work. None of them is exotic, which is precisely the point.

1. Leadership ran it as a programme, not a memo. There was a start date, a cadence of sessions, and an end date with a measurement attached. Most companies' AI strategy is an email encouraging people to “explore the tools.” A programme has an owner and a calendar; a memo has neither.

2. Reps on real work, not tool demos. The proficiency that moved was proficiency at redesigning their own workflows. You do not get a 41-point jump in workflow automation by watching someone demonstrate a chatbot. You get it by taking the report you actually write, the pipeline you actually manage, the process you actually own, and rebuilding it with the model, repeatedly, with feedback.

3. A champion in every department. The moment someone gets stuck, the distance to help determines whether they push through or quietly revert to the old way. A peer three desks away who has already solved a similar problem beats a central help channel every time. Champions also translate: the marketing champion talks about marketing workflows, the finance champion about finance ones.

4. Measurement bookended the quarter. The baseline survey forced an honest reckoning with where the team actually was, and the promise of a re-survey created accountability all quarter. Measurement is also what let them publish a result instead of a vibe. Most companies cannot tell you whether their AI training worked because they never defined what working would look like.

5. Nobody was exempt. The fastest way to kill an AI programme is for the team to notice that the senior people treat it as optional. Universal participation is not about fairness; it is information. It tells every employee this is how the company works now, not a side project that will blow over by Christmas.

Why the Typical Rollout Fails

Now compare that with the AI rollout most UK businesses have actually run. It usually looks like this: leadership buys licences, someone books a lunch-and-learn, an enthusiastic presenter shows some impressive demos, everyone nods, and then everyone goes back to their inbox. There is no second session. There is no practice on anyone's real work. Nobody owns adoption as a responsibility. Nothing is measured, so three months later the only data point is the licence-usage report, which is quietly embarrassing, and the conclusion drawn is that “our team isn't really AI people.”

The team was never the problem. A single exposure with no reps, no support structure and no accountability would fail to teach anyone anything: Excel, a new CRM, a language. AI is not different; it just gets rolled out worse, because the demos are so impressive that leadership mistakes exposure for training. Acquisition.com's quarter is useful precisely because it shows what the ordinary machinery of skill-building (sessions, practice, coaching, measurement) does when someone bothers to apply it to AI.

Want this quarter run for your team? See our AI team training

The UK SME Version of That Quarter

You do not need Acquisition.com's budget to run their playbook. A 15-person accountancy practice, a 40-person manufacturer, a 100-person services firm: the structure scales down cleanly because none of its parts is expensive. Here is what a focused quarter looks like for a UK SME.

Week 1: baseline. Survey the team on the same handful of skills: prompting, redesigning workflows, validating outputs, judging when AI is worth using. Alongside it, inventory the ten most repetitive workflows in the business: month-end reporting, proposal drafting, debtor chasing, meeting follow-ups. This week costs almost nothing and changes everything, because it converts “we should do something about AI” into a list and a number.

Weeks 2–5: structured sessions on your actual workflows. A session every week or fortnight, each one built around workflows from your inventory, not generic demos. The team leaves each session having built something they will use on Monday. This is the phase where an external trainer earns their fee: our live AI workshops (from £2,500) are designed to be exactly this: your workflows, your data formats, your team's hands on the tools.

Throughout: weekly office hours and named champions. A standing 45-minute slot where anyone brings a stuck workflow, plus one named champion per team or department, usually your most curious person, not your most senior. Their job is not to be an expert; it is to be the nearest unblocking point and to collect what is working.

Weeks 6–11: embed and extend. One workflow per fortnight moves from “we tried it in the session” to “this is how we do it now,” with a documented prompt and a named owner. Individuals who want to go deeper can run a self-paced course in parallel, our AI for Finance course (£99) covers the finance-specific reps.

Week 12: re-survey and decide. Run the same survey, compare the numbers, count the workflows now in live use, and pick the next quarter's targets. If the numbers moved, publish them internally: nothing sustains a programme like visible progress. This structure is not theoretical for us: our finance team training case study walks through what happened when a real UK finance team ran a version of this cycle.

The Three Questions to Answer First

Before the quarter starts, a business owner needs three questions answered, because these are the three places the typical rollout dies.

What should we prioritise? Not every workflow is worth automating first. The right first targets are high-frequency, format-driven and low-risk: the work that repeats every week in a known shape. The baseline-week inventory answers this; without it, teams default to whatever the last demo showed them.

Who will help us use it? Tools without a support structure produce a spike of curiosity and then silence. The answer is the combination that worked at Acquisition.com: a trainer for the structured sessions, office hours for the stuck moments, champions for the everyday questions. Our team training programmes supply the first two and help you set up the third.

How does it become business outcomes? Proficiency scores are the leading indicator; hours saved and cycle times are the lagging one. The bridge between them is the workflow list: every skill gained in a session should attach to a named workflow with a named owner, so that by week 12 you are counting not just confidence but working automations. That is the difference between a team that did an AI course and a team that runs differently.

And the honest summary of the whole announcement is this: the companies you assume are years ahead are one focused quarter ahead, and they have handed you the recipe. The gap between your team and an AI-first team is not talent, budget or timing. It is one measured, structured, nobody-exempt quarter, and the best week to start it is this one.

Frequently Asked Questions