The Mobility Trap: lateral career moves plunge across 57% of firms as internal pathways collapse.

LDNUCU · 27 October 2026
Course completion is not capability.
As of October 30, UK employers face stricter statutory duties under the Worker Protection and Employment Rights Act requiring the proactive prevention of sexual and third-party harassment. It is a substantial shift. If your organization has rolled out a mandatory compliance video, collected digital signatures and considered its legal exposure mitigated, the new regulatory environment is about to expose a critical flaw in that logic.
The October deadline is a lens on something wider: the dangerous gap between deploying a tool or a policy and building human capability. The same gap appears in digital harassment, in complex construction contracts, in the rollout of artificial intelligence and in the quiet closing of internal career pathways.
Listen first
The full edition, as a podcast episode.
The thread through this edition
Four assumptions to test
Whatever the industry, the first principle is the same. A 100% completion rate on a static harassment module does not mean workplace behavior has changed, just as logging into a new software platform does not mean your workforce knows how to use it. Press each pair and see what the reporting says about it.
A 100% compliance completion rate on a static harassment module does not mean workplace behavior has actually changed.
Giving an employee a login to an enterprise AI platform does not make them capable of using it.
A polished video on teamwork means little if a site manager cannot administer a collaborative contract when a critical material shipment is delayed by three weeks.
Generating a hundred microlearning courses with an AI authoring tool does not mean the workforce learned anything.
Compliance
The checklist assumed an office
Pre-2021 checklist policies were designed for a physical, geographically bound workplace. They implicitly assumed harassment happens by the water cooler or in a conference room: places with witnesses, physical context and standard working hours. The permanent shift to hybrid and remote work has pushed employee interactions into what the reporting calls the digital gray area.
A legal or HR team might argue that a policy is a policy. If the company updates the digital handbook, emails new rules of conduct for Slack and Zoom and collects an “I agree” click, surely that covers the company if an employee goes rogue in a chat thread? Regulatory bodies and tribunals are explicitly rejecting that logic. The statutory mandate requires employers to take all reasonable steps to proactively prevent harassment. A passive digital signature on a PDF does not prove an employer took reasonable steps to change behavior.
Harassment today is rarely a shouted insult in a hallway. It is far more subtle, and a checklist does not account for the psychological reality of sitting in your living room, trying to interpret an ambiguous message from someone who holds power over your career progression.
A shared location
Witnesses
Physical context
Standard working hours
Four conditions the old checklist took for granted. Select a moment below to see which one it removes.
Each moment removes one condition the checklist relied on. The signature stays exactly as it was.
1,294
sexual discrimination tribunal claims were filed in the fourth quarter of 2025 alone.
Recent updates from Ciphr, which tracks HR compliance data, and warnings from Gi Group UK on tribunal claims.
The penalties being tracked reach up to £500,000. Average individual awards sit between £27,000 and £45,000. There is also a new 25% compensation uplift if an employment tribunal finds that an employer breached its preventative duty.
At the lower end of the reported average range, £27,000 becomes £33,750 with the uplift. At the upper end, £45,000 becomes £56,250.
Illustration only: the 25% uplift applied to the reported average range. It is not a forecast of any individual award.
What a reasonable step looks like
Practice in the gray area
Distributing a PDF about road safety does not make someone a defensive driver. Yet that is how many organizations treat digital conduct: they assume the workforce can navigate ambiguous interpersonal dynamics online because it read the rules. If the traditional approach is broken, building the capability takes practice that simulates the digital friction employees actually meet.
The quiz everyone passes
This is the approach the reporting calls into question: a multiple-choice quiz that everyone passes, built around an obvious example rather than the ambiguity employees actually face.
The gray-area scenario
Instead of spotting misconduct that is easy to spot, the learner is placed in a branching scenario and has to decide. The reporting describes the questions; it does not supply a single right answer to them.
The manager who intervenes
A project channel. One teammate is being quietly frozen out.
Managers need to recognize the subtle signs of digital exclusion, such as a team member being quietly frozen out of a project channel, and be trained to intervene effectively. That makes them active guardians of workplace culture rather than passive administrators chasing annual compliance sign-offs.
That shift from passive administrator to capable guardian is where organizations stumble. The failure to treat operational risk as a formally trained capability is not confined to HR compliance. It is structurally identical to a challenge in a very different sector, where the same assumption of competence stalls major infrastructure projects.
Construction and engineering
The contract is the operating system
Research from NEC Contracts across the UK construction and engineering industry describes a similar capability failure. Across major infrastructure investments, cost overruns hit 43% of projects, delays affect 37% and commercial disputes hold up another 37%. The root cause rarely comes down to a lack of engineering talent or poor raw materials. It points back to a substantial gap in commercial capability.
NEC contracts set the standard for how global infrastructure projects are procured and managed, and they require a specific administrative skill set. Yet commercial contract management is widely treated as an informal site habit rather than a formally trained capability: shadow a colleague for a week and hope to pick it up.
In modern construction, a collaborative contract is not a legal document that is signed and stored in a drawer until litigation begins. It is the operating system for the daily project. It dictates how early warnings of delay are managed, how compensation events are calculated when supply chain issues hit, and how commercial risk is shared between contractor and client.
Learned by shadowing a colleague
Early warnings of delay
Compensation events calculated
Commercial risk shared
Trained as a capability
Early warnings of delay are managed
Compensation events are calculated
Commercial risk is shared between contractor and client
Both views illustrate the argument made in the reporting. The research itself measures overruns, delays, disputes and training investment, not the effect of any one training method.
Here the editorial lens applies directly: engagement is not performance. A site manager can complete an interactive walkthrough or watch a polished video on teamwork, but if they cannot administer a collaborative contract when a critical shipment is three weeks late, the business impact is lost. It is like handing someone the blueprint to a skyscraper and expecting them to pick up structural engineering on the job.
Modern construction contracts call for real-time legal literacy, digital documentation skills and proactive conflict resolution. Leaving that to osmosis does not build capability. It generates costly dispute arbitration months down the line. Training has to reflect the operational reality of the job, which means high-pressure commercial decision-making alongside technical engineering.
Artificial intelligence
A grand piano is not an orchestra
If mature industries fail to train professionals formally on collaborative commercial frameworks, how are organizations handling the rapid rollout of tools like artificial intelligence? The reporting suggests they repeat the same errors faster: a systemic misalignment between technology investment and human enablement.
A survey of senior leaders by Kaplan, alongside data from Learning News, finds that 37% of leaders report missed financial returns on their AI investments. They do not blame the technology. They point to internal skills gaps. Budgets point the same way.
of senior leaders report missed financial returns on their AI investments, and cite internal skills gaps as the primary cause.
of organizations spend more on AI software licenses than on employee enablement.
of employees report being instructed to use AI in daily tasks without any structured training.
Kaplan survey of senior leaders and Learning News data. Each grid shows 100 organizations or employees.
What share of the workforce do HR and L&D leaders judge genuinely proficient with AI?
of employees report feeling confident using AI.
is what HR and L&D leaders estimate to be genuinely proficient, testing the ability to produce usable, accurate output.
The two figures measure different things: how confident employees feel, and whether they can generate usable, accurate output as assessed by HR and L&D leaders. The gap between them is 43 percentage points.
A significant perception gap hides this from the executive suite. Vendor marketing frequently feeds the disconnect by selling the idea that an AI seat license automatically produces instant productivity. The governing rule stays the same: tool adoption is not enablement, and content production is not learning. Buying every employee a grand piano does not produce an orchestra. Without training, it produces expensive noise, which in a company looks like error rates, rework and hallucinated data.
True AI value does not come from querying a chatbot with a generic prompt. It requires task restructuring. A knowledge worker has to break down a daily workflow, identify which specific nodes AI can reliably handle, apply critical prompting and use domain-specific judgment to validate the output. You have to know what an accurate, high-quality answer looks like to recognize when the system is confidently hallucinating. The reporting describes employees who submit AI-generated reports citing non-existent legal precedents, or inventing historical details, because they lacked the expertise to detect the error.
The task is restructured
Told to use AI, no structured training
When employees are required to use AI without structured training, productivity does not automatically rise. They spend two hours crafting ineffective prompts, get unusable or generic output and quietly rework the task from scratch to meet the deadline. That is effectively doing the work twice, and it adds invisible friction to the daily workflow.
Talent and mobility
The Mobility Trap
Organizations demand that the workforce adapt to AI and redesign complex workflows. At the same time, market signals show companies restricting the internal pathways needed to do it.
Fosway Group’s 2026 9-Grid for Talent & People Success found that 40% of HR teams face severe economic headwinds that translate into significant budget pressure. Combined with global workforce data from Workday, the picture is that lateral internal job mobility has dropped at 57% of firms surveyed. More than half of employers have seen internal career moves decline.
For the professionals who build organizational capability, the signals are no kinder. Workday’s data indicates that job requisitions for Learning and Development professionals fell by 13%, while mentions of management skills in job postings declined by 7%.
of business leaders expect substantial AI productivity gains from existing staff
40%The goal: a job redesigned around AI
Three supports are missing between today’s job and a redesigned one. Restore them in order and watch whether the climb can continue.
The gaps are symbolic. The three measures describe different things and cannot be added together, and the reporting does not measure what restoring them would produce.
That produces a paradox. 40% of business leaders expect substantial AI productivity gains from existing staff, seeking higher output and innovation, yet organizations cut the very functions responsible for guiding the change. They want the destination without investing in the road.
Freezing internal mobility traps employees in legacy workflows. Cutting L&D professionals and de-prioritizing management capability leaves them isolated with technology they do not fully understand. The result is talent stagnation, and stagnation breeds a disconnected workforce, the opposite of the agile organization leadership aims to build. The assumption that technology alone drives productivity without human development is an expensive illusion.
The lesson across all four stories
Technology is the amplifier. Human capability is the signal.
Whether the subject is the statutory duty to prevent digital harassment in hybrid workplaces, the administration of major infrastructure contracts or the enterprise rollout of AI, the lesson is the same. Organizations cannot buy their way out of capability gaps with software licenses or policy checklists. None of these systems works without human capability.
Without building capability, an organization merely amplifies noise and operational friction: error rates, rework and hallucinated data.
Solid line: the output. Dotted line: the signal alone. No measurements are shown.
Are you funding the illusion of capability by purchasing tools and tracking completion rates, or the practice and development that change workplace behavior?
Organizational resilience takes sustained investment in human judgment and management capability alongside technical tools. Look at your organization’s budget this quarter and ask which one it is paying for.