Estimated reading time: 10 minutes
Takeaways
- Treat upbeat headline data, PMI included, as a question, not a verdict. Check what’s driving it first.
- Check leadership belief against floor data before you act on it. The two are drifting further apart than most assume.
- AI’s real barrier isn’t culture any more. It’s data quality and unclear ROI, so fix the data before you scale the pilot.
- Skills shortages and rising costs aren’t waiting for policy clarity. Your automation decisions shouldn’t either.
- Build one small feedback loop, a question, a log, a weekly check, before you reach for a new system.
- Treat H2 2026 as the half for testing assumptions, not the half for launching programmes.
H1 2026 didn’t hand manufacturing a new problem. It handed us reports that finally measured an old one and wouldn’t let it stay anecdotal. The real question for H2 isn’t whether the gap exists. It’s what you’ll find once you check your own patch.
Six months ago you likely believed your ops team had a real say in decisions. Four new reports say think again. They cover AI readiness, leader sentiment, SME policy, and shop floor reality. All four land on the same gap. What leaders believe is true. What the data shows is true. The two have drifted apart.
That’s the real story of H1 2026. It’s not a new crisis. It’s old gaps, finally counted, all at once, by people who weren’t talking to each other.
The manufacturing intent-execution gap is the distance between what leaders believe is happening on the shop floor (involvement, data trust, AI value, policy benefit) and what staff and operational data actually report. It is distinct from simple miscommunication: leaders are not withholding information, they are working from a genuinely different vantage point that several 2026 UK manufacturing reports now measure as wide as 47 percentage points.
The reality nobody planned for
Start with the number that should make this easier. The UK Manufacturing PMI hit 53.9 in May, its best reading since May 2022, with seven straight months above the 50-point growth line [1]. On paper, the sector is bouncing back. Look closer and it wobbles. Much of that strength is firms front-loading orders and stockpiling, bracing for price rises and supply trouble, not fresh demand [1]. Input costs are rising regardless. Oil-driven fuel prices jumped 55.6% in the year to May, after war hit the Middle East in March, and the pound has slipped too [2]. Even the good news comes with a footnote.

Here’s the constraint sitting on every ops leader’s desk right now. Employment costs are rising again, for the second year running, and 86% of firms expect more of the same [3]. The Living Wage climbs 4.1% to £12.71 an hour from April. Tax relief on equipment just got messier too: a new 40% first-year allowance lands, but the older writing-down rate was cut [3]. Meanwhile 48,000 manufacturing jobs sit unfilled. Skills shortages are the top hiring barrier for nearly three in four firms [3].
Now add a workforce that doesn’t trust its own data. Operations Feedback Systems (OFS) surveyed 500 UK manufacturing staff. Only 11% fully trust the data they use [4]. Two in three use quiet workarounds instead, because the system misses the real cause [4].
That mistrust costs real money. 46% of staff have seen downtime linked to slow visibility. 41% say it pushed costs up. 33% missed a deadline because of it [4]. These sit on your P&L.

So you’re juggling rising costs, a skills gap, and a trust gap, all in six months. None of that is new. What’s new is how clearly it’s now on record.
The insight: belief and reality have quietly split
Here’s the pattern that runs through all four reports. It’s worth sitting with.
95% of C-suite leaders believe their ops teams are genuinely involved in decisions. Only 48% of mid-level staff agree. Only 51% of entry-level staff agree. That’s a 47-point gap [4].

The same shape turns up in AI. Cultural pushback against AI is now the smallest barrier on the list, just 3.4% [5]. Nobody’s arguing against the idea any more. Yet almost one in five firms still say AI’s value is unclear inside their own walls. Over 70% call themselves cautious followers, not leaders, on AI [5]. Belief in the tech has outrun proof that it works here.

Policy shows the same split. 63% of manufacturers think the Industrial Strategy will lift their investment plans [3]. Ask SMEs themselves, and only 10% believe it will help this year. 75% feel barely heard, or not heard at all [6].
Track the Modern Industrial Strategy policy story over three quarters and it sharpens further. Government’s own updates reported over £250bn in investment commitments in Q3 2025, then £79bn in Q4, then £35bn in Q1 2026 [7][8]. The figure swung hard. The tone never did. Every update reads the same: steady delivery, no slip ever admitted. The SMEs these numbers are meant to help haven’t moved either. Still 10% confident, still 75% unheard, whichever quarter you check.

Four domains, now. Macro signals, cost, tech, policy. The same gap, every time: belief at the top, doubt at the level where the work gets done. None of this means leaders are lying to themselves on purpose. It means the view from the top floor is just a different view. It was never the full picture. It just felt like one, until someone went and asked.
A note on method
PMI, producer prices, and the investment figures above are hard, official, monthly numbers. The survey findings sitting alongside them are self-reported, and vary in sample and method. They’re paired here on purpose, not because they’re statistically equivalent, but because the macro data tells you what’s happening across the sector, and the floor data tells you how it feels to live through it. Read each domain on its own terms, then ask which one your own operation is actually living in.
A solution lens for the second half
H1 didn’t hand us a fix. It handed us a habit worth building. Try this:
- Pick one metric your floor team distrusts or works around. Ask them why.
- Test one leadership belief (about involvement, AI value, or policy benefit) against what your own people report. Expect a gap.
- Choose one AI use case with clean data behind it. Quality control and maintenance score highest for impact, start there [5].
- Add five minutes to your existing shift handover for one piece of floor-level input. Don’t book a new meeting.
- Check, after two weeks, whether that input changed a single decision. If not, you’ve just found your real starting point.
- Repeat with the next belief. Don’t wait for a strategy document to give you permission.
What this looked like in practice
A plant manager at a mid-sized fabrication firm thought his shift teams felt heard. He held an open-door policy and monthly briefings. Then he asked his shift leads how decisions actually got made, and the answer surprised him: most issues got solved quietly on the floor and worked around, because the system never caught the real cause. He didn’t buy a new system. He added one habit: every handover ends with one question, logged in a shared sheet his ops director checks weekly. Three months on, two recurring downtime causes finally surfaced.
Common mistakes still being made
- Treating leadership confidence as proof of floor-level reality, instead of checking it against the floor.
- Designing new systems without floor input before rollout. Only 18% of firms do this [4].
- Launching AI pilots before fixing the data quality issue that 19% of firms already flag as their top barrier [5].
- Waiting for policy clarity before acting, when 73% of SMEs are already investing in automation regardless [6].
- Reading low pushback as high trust. Cultural resistance to AI sits at just 3.4%, but that doesn’t prove the value case. It just means nobody’s arguing [5].
What good looks like
You’ll know the gap is closing when floor reports change a call within a fortnight, not a quarter. When your AI use cases fit on one hand, each tied to a named metric. When handovers surface workarounds instead of hiding them.
None of this needs a big launch. It needs a small loop that runs every week, without fail. Most firms don’t lack the will. They lack a habit that survives the first busy week.
What to do on Monday
- Ask three shift leads, separately, whether they feel involved in decisions affecting their line, and compare the answers to what you assumed.
- Identify one workaround your team uses instead of the official reporting system, and find out why it exists.
- Pick the single AI use case with the cleanest, most available data in your operation and scope it narrowly.
- Add one specific question to your existing shift handover this week rather than booking a new meeting.
- Review what changed after two weeks. If nothing did, that gap is your real starting point for H2 2026.
What is the manufacturing intent-execution gap?
It is the measurable difference between leadership belief about operational reality, such as staff involvement or AI readiness, and what frontline data and staff actually report. UK research in 2026 found a 47-point gap between C-suite and frontline views on whether operations teams are meaningfully involved in decisions.
How is this different from a simple communication problem?
Communication problems are usually solved by saying more. This gap persists even where leaders hold regular briefings and open-door policies, because the official systems collecting operational data often miss the real cause of floor-level issues. The fix is structural, not just more talking.
What is a practical threshold for spotting this gap in my own operation?
A useful signal is reliance on workarounds. If two in three of your operations staff regularly bypass official systems, as UK research found in 2026, that is a strong indicator your reported data does not reflect floor-level reality.
Who is responsible for closing this gap?
Operations leadership owns it, because only leadership can change how decisions are tested against frontline input before they are made. Frontline staff cannot close a gap they did not create and do not control the systems behind it.
Sources
Sources below mix official statistics, fielded surveys, and one working draft (marked). They’re not weighted as equal evidence, see “A note on method” above.
- S&P Global / CIPS, “UK Manufacturing PMI”, May 2026 release
- Office for National Statistics, “Producer Price Inflation, UK: May 2026”
- Make UK and PwC, “Executive Survey 2026: Time for Mission Growth”
- Operations Feedback Systems (OFS), “Same Factory, Different Reality” 2026
- The Manufacturer, Oracle and IBM, “The Manufacturing AI Readiness Index 2026”
- Made in Group, “Priorities for UK Manufacturing SMEs 2026”, working draft
- UK Government, “The UK’s Modern Industrial Strategy: Quarterly Update, July to September 2025”
- UK Government, “The UK’s Modern Industrial Strategy: Quarterly Update, October to December 2025”
Discover more from Nick Leeder & Co
Subscribe to get the latest posts sent to your email.

