
Outlook
Part of Web design agencies outlook: UK evidence limits
Web design agency skills in England: GOV.UK service team roles, six editorial 2027 scenarios and a capability register
Which skills to look for in a web design agency in England: GOV.UK service team roles today, six editorial 2027 scenarios and a capability register for briefs.
What to take away
- Today's skills and the 2027 scenarios stay apart. The first come from a published roles page, the second are editorial scenarios with no probability.
- No England agency vacancy count, salary series or headcount estimate is given, because none is verified.
- Plan around demonstrable capabilities, then recheck demand against real briefs.
- The published roles show capacities to test in procurement, not fixed job titles.
- Review the capability register when scope, platform, data, supplier or guidance changes.
Current web-delivery skills, named by source
The GOV.UK Service Manual page on service team roles lists the roles or skills a team building a government service needs in the team or available to it: product manager, service owner, delivery manager, user researcher, content designer, designer and developer.
The same page names other roles a larger service may need: performance analyst, technical architect, DevOps engineer, business analyst, quality assurers and testers, and technical writers.
The population is government teams. The list does not prescribe a private agency's titles, staffing ratio or organisation chart. It also notes that the skills a team needs change throughout the lifecycle of a service, so a role list is a starting point for a brief, not a fixed structure.
A user researcher helps the team learn about the people who will use the service. A content designer is responsible for the content. A developer builds accessible software with a focus on what users need. A delivery manager makes sure accessibility is factored into each feature or activity. A performance analyst collects and presents key performance data.
Quality is the responsibility of the entire team, and the final responsibility lies with the service owner. For a buyer, that means a brief should say who owns quality, not only who writes code.
Accessibility as a checkable capacity
Government accessibility requirements give a buyer something concrete to test. Services must meet level AA of the Web Content Accessibility Guidelines 2.2 as a minimum. They must also work on the most commonly used assistive technologies, including screen magnifiers, screen readers and speech recognition tools, and carry an accessibility statement.
Meeting them also meets the accessibility regulations for public sector websites and mobile applications, whose full name is The Public Sector Bodies (Websites and Mobile Applications) (No. 2) Accessibility Regulations 2018.
The requirements are set out in the Service Manual's accessibility introduction, written for government services. A private project needs its own applicable standard, user tasks and qualified review, but the same checks work as procurement questions: which standard, which assistive technologies, which statement.
Secure delivery as a capacity
Ask agencies that offer AI-assisted work about four stages of secure AI system development (NCSC guidelines for secure AI system development): secure design, development, deployment, and operation and maintenance. Security must be a core requirement throughout the life cycle.
Those stages are not a certificate for a worker or a course. They give a buyer four headings to ask about when a supplier proposes AI-assisted work. The risk scenarios show how to turn such questions into delivery controls.
Six editorial scenarios for 2027
The table pairs each current skill with an editorial scenario item for 2027. The scenarios carry no probability. They name who owns the capacity and what evidence a buyer can check.
| Current skill, from the roles page | 2027 scenario item | Scenario owner | Evidence a buyer can check |
|---|---|---|---|
| User research | accessibility findings on a live service | researcher and test lead | scoped evidence for representative sampling |
| Content design | AI-assisted copy with source tracing | content designer | traced sources and rights |
| Interaction design | interface change with retained test evidence | developer | accessible test evidence |
| Development | release stop and restore authority | delivery lead | last approved state restored |
| Quality assurance | model-change notices in tenders | buyer and supplier | conformance evidence at model change |
| Performance analysis | export or deletion evidence on request | security owner | export or deletion record |
Editorial scenario: buyers ask teams to pair production skill with assurance. To test it, track these signposts from the market outlook in a defined buyer sample: mandatory tender artefacts, failed acceptance gates, incident reviews and model-change notices. If comparable briefs do not ask for these, the scenario is disconfirmed.
Reading an agency proposal against the roles
A web design agency in England may be a design studio, a development agency or a full-service digital agency. Each covers the roles above differently. The table shows what to look for in a proposal and which evidence to ask for. It names no agency and ranks none.
| Agency type | Roles the proposal should name | Evidence to request |
|---|---|---|
| Design studio | designer, content designer, user researcher | research plan, content samples, accessible prototype |
| Development agency | developer, technical architect, tester | test approach, release and rollback process |
| Full-service digital agency | delivery manager, service owner, performance analyst | named owner of quality, reporting sample |
If a proposal leaves a role empty, ask who covers it and how that person will be paid for the time. A role covered by "the team" has no accountable owner. Public-sector briefs, such as those published on Contracts Finder, show how buyers phrase these capacities.
Build a capability register
For each brief, list the decision or task, the required competence and the accountable owner. Add acceptable evidence, independence need and current capacity. Then record the gap, the learning or procurement route, and a review date. Include a manual or existing-team baseline before adding a role.
Here is a filled example for an illustrative brief: a small charity website redesign. The rows are invented for the method, not taken from a real agency or tender.
| Decision or task | Competence | Accountable owner | Acceptable evidence | Gap and route |
|---|---|---|---|---|
| Approve page templates | Interaction design | Designer | Tested prototype with notes | None |
| Check WCAG 2.2 AA | Accessible development | Developer | Independent test report | Buy an external audit |
| Edit content after launch | Content design | Content designer | Editor trial with staff | Train two staff |
| Restore after a failed release | Release control | Delivery manager | Restore test record | Add a rollback step |
Take a public-sector website brief published on Contracts Finder. It states the buyer's requirement and the deadline. Map each stated requirement to an accountable owner and the evidence a buyer can check. The selection checklist shows how to record the result at each gate.
The team roles article sets out the same capacities without job titles. Check the market guide for the delivery context.
Review the register when scope, platform, data or supplier change, and when applicable rules change. Stop delivery when a mandatory capacity has no competent owner.
Before you act
- List the decision or task for each brief.
- Name an accountable owner and acceptable evidence.
- Include a manual or existing-team baseline before adding a role.
- Record the accessibility standard and assistive technologies to test.
- Set a review date for the register.
Common questions
Can a skills forecast be built from vacancy pages?
No. A forecast cannot be derived from a few vacancy pages. Plan around demonstrable capabilities, then recheck demand against real briefs.
What do the GOV.UK roles show a private buyer?
They show capacities to test: research, content, accessible development, delivery and performance analysis. They do not set job titles or staffing ratios for a private agency.
How should AI use be handled in a capability register?
AI use needs human evaluation, data and rights judgement, and security ownership. Ask the supplier how secure design, development, deployment and operation are covered.



