Nobody embraces change. They agree to try it

The word embrace is doing a lot of concealing. It proposes that the goal is enthusiasm, that enthusiasm is produced by communication, and that a person who has not produced any is a problem for the communications workstream.

What people actually do is agree to try something, on terms. Two of the three terms they care about are already written down in British statute, with clocks attached, and most transformation plans schedule their engagement activity after those clocks have run out.

The redundancy clock is a delivery milestone

Section 188 of the Trade Union and Labour Relations (Consolidation) Act 1992 requires an employer proposing to dismiss as redundant 20 or more employees at one establishment within a period of 90 days or less to consult appropriate representatives. Consultation must begin in good time, and in any event at least 30 days before the first dismissal takes effect, or 45 days where 100 or more dismissals are proposed. It must cover ways of avoiding the dismissals, reducing the numbers to be dismissed and mitigating the consequences, and it must be undertaken with a view to reaching agreement.

The part that catches technology programmes is the definition. Section 195 provides that references to dismissal as redundant are references to dismissal for a reason not related to the individual concerned, and that where an employee is or is proposed to be dismissed it shall be presumed, unless the contrary is proved, that the dismissal is by reason of redundancy.

Nobody has to say the word redundancy for this to bite. A programme that redesigns roles, and results in twenty people leaving for reasons that are not about those twenty people, is inside the section, and the presumption runs against the employer.

Read the consultation content requirement again with a delivery plan in front of you. It requires consultation on ways of avoiding the dismissals. A consultation that opens after the target operating model has been signed off, the system has been configured to it, and the go-live date has been announced to the market cannot honestly consider avoidance, because avoidance is no longer available. The statute is not asking for an update meeting. It is asking for a period in which the outcome is genuinely open.

So the sequencing is a constraint on the plan, not a compliance task appended to it. Any transformation whose people impact crosses that threshold has a legal event that must start weeks before go-live and must be capable of changing the design. Programmes that discover this late do one of two things, and both are expensive: they slip, or they proceed and buy the risk.

For a public body, an unusable system is a legal failure

The second term is accessibility, and in the public sector it is not a matter of good practice.

Regulation 6 of the Public Sector Bodies (Websites and Mobile Applications) (No. 2) Accessibility Regulations 2018 requires public sector bodies to comply with the accessibility requirement, subject to the disproportionate burden assessment in regulation 7. Regulation 12 then makes the consequence explicit: a failure to comply with the accessibility requirement is to be treated as a failure to make a reasonable adjustment, by reference to sections 20, 21 and 29 of the Equality Act 2010.

Intranets and extranets are in scope, with a carve-out for content published before 23 September 2019 that lasts only until the site undergoes a substantial revision. Replacing an internal system is a substantial revision.

The practical effect is that a colleague saying the new system is harder to use may be describing a duty rather than expressing a preference, and the organisation’s response to that report is a legal position rather than a change management stance. Very few internal transformation programmes treat accessibility of staff-facing systems with the seriousness they apply to public-facing ones, and the regulations do not draw that distinction.

There is a second-order effect that programme managers underrate. The regulations also require an accessibility statement, and they provide a route by which a person may request information in an accessible format, with a failure to respond satisfactorily treated the same way. So the position is not simply that an inaccessible system creates exposure. It is that the organisation has to be able to say, in writing and in public, what does not work and why. An accessibility statement drafted honestly during a rollout is the most useful document a transformation produces, because it forces someone to write down the parts of the design that were left for later, and later is where they usually stay.

The third term is what happens when the system is wrong

The first two are written down. The third is the one an organisation has to write for itself, and it is the one that decides whether the other two hold.

Britain has the most thoroughly documented case in the world of what happens when a large organisation treats a user’s report about a system as information about the user. Sir Wyn Williams published Volume 1 of the Post Office Horizon IT Inquiry’s final report in July 2025, dealing with redress and with what he describes as the disastrous human impact of the scandal. The Inquiry has held 226 days of hearings, taken oral evidence from 298 witnesses, and had around 274,604 documents disclosed to it. Among the matters it set out to examine were the handling of reported discrepancies, meaning how the Post Office responded when postmasters reported accounting shortfalls, and the institutional culture that contributed to what followed.

Hundreds of postmasters, managers and assistants were wrongly accused of theft and fraud because of faults in an IT system. The recommendation in Volume 1 that reaches furthest beyond the Post Office is the seventeenth, that government establish a standing public body to create, administer and deliver financial redress schemes for people wronged by public bodies. It exists because the mechanism failed at scale and there was nothing standing to catch it.

No comparable organisation intends this. What every organisation running a system that produces figures people are held to has available is the same failure mode: a report from a user that the numbers are wrong, arriving at a function whose working assumption is that the numbers are right.

The term worth offering, therefore, is procedural and cheap. A route by which a user’s report that the system is wrong is recorded as evidence about the system, with a named owner, and a count of open reports that appears in the same pack as the delivery milestones. Not a helpdesk ticket queue, which measures resolution. A count of unexplained discrepancies, which measures doubt, and which somebody senior has to look at while the programme is still running.

What people are actually negotiating

Three things, in this order.

What happens to my job. Governed by statute where the numbers cross a threshold, and it has a clock.

What happens to how I am measured. Almost never decided, and usually the real objection. A change that alters what a person does while leaving their target where it was is a request that they absorb the transition out of their own performance rating. Anybody would refuse that, and refusing it is not a cultural problem.

And what happens when the system is wrong and I say so. This is the one that determines whether the first two mean anything, because a person who believes they will be blamed for the system’s errors has correctly identified that the terms are not being honoured.

None of that requires enthusiasm. It requires an offer with dates and named owners. Communication campaigns run in place of an offer because an offer costs something and a campaign does not, which is also why the campaigns do not work.

What happens when nobody makes that offer, and a rollout meets a department that never asked for it, is documented in pilots clear, rollouts do not. Everything else this desk publishes is collected under digital transformation.