DWP published a commercial notice on 25 March inviting tech suppliers to help shape procurement plans for the new Jobs and Careers Service. The department has £55m to spend this year on the merger of Jobcentre Plus and the National Careers Service, with £20m earmarked for digital activity.

If your charity delivers DWP-funded employment support, this notice is worth reading carefully. It tells you what the department plans to automate, where it sees gaps, and, by implication, where human-delivered services will need to prove they're still worth funding.

The 12 categories

The notice lists 12 areas where DWP wants to understand what technology and AI products are available. They are: career advice, career planning, skills-gap analysis, financial planning, career coaching, learning and training provision, job discovery, job market insights, CV support, job application support, application tracking, and interview support.

Between now and the end of September, DWP will run a programme of market engagement. The first event is a two-hour Microsoft Teams briefing on 2 April. After six months of engagement, the department will decide how to proceed with formal procurement. It may run multiple procurements covering individual categories or groups of categories, in parallel or in phases.

Suppliers can register interest in some or all of the 12 categories. DWP has said no preference will be given based on how many categories a supplier covers.

What this means for charities

The NCS contracts expire on 30 September 2026. DWP is bringing the careers service in-house, with roughly 1,000 careers advisers expected to TUPE into the civil service. The sub-contractors who currently deliver much of the NCS at local level are losing a major income stream. Some smaller organisations may not survive it.

At the same time, DWP is procuring AI tools that can do at scale what many charities currently do by hand: write CVs, match people to vacancies, identify skills gaps, prepare people for interviews. The stated objective of JCS is a "digital, universal and fully inclusive service" with "high-quality personalised support." For people not on benefits, the department has said support will be "served digitally, through a self-service option."

None of this means charities are finished. But it does mean the argument for funding human-delivered employment support has to change. If DWP can offer automated CV writing and job matching to anyone with a phone, a charity pitching "we help people write CVs" is in trouble. The value of human support is in the things automation can't do: building trust with someone who doesn't trust institutions, navigating complex barriers that don't fit a dropdown menu, and working with people whose needs are too specific or too messy for a general-purpose digital tool.

Where the gaps are

Read the 12 categories again. Notice what's missing.

There is nothing about language barriers. Nothing about ESOL. Nothing about cultural readiness for UK workplaces. Nothing about the specific needs of refugees, people with limited literacy, or anyone whose first barrier to employment isn't "finding a vacancy" but "being able to communicate in the interview."

An AI career coaching tool that generates a skills-gap analysis is useless for someone who can't read the output. An automated CV builder doesn't help someone who doesn't know the English word for what they did in their previous job. Job discovery tools are irrelevant to someone who can't understand the job description.

DWP knows this. The Get Britain Working white paper talks about "real time language translation" in innovation hubs. Connect to Work, the £1.2bn successor to Restart, explicitly includes refugees. But the commercial notice focuses on the digital services that will work for the majority of jobseekers, not the hardest-to-help cohorts. That's where charities come in.

What you can do this week

If you manage a programme at an employment charity and you're wondering how JCS affects your funding model, here are four things worth doing before the April briefing.

First, map your current service against the 12 categories. Which of those 12 things do you do today? For each one, ask honestly: could a digital tool do this as well as we do, for most of our clients? If the answer is yes, that part of your service is at risk. If the answer is no, write down why not, in specific terms. "We work with refugees who have limited English and complex immigration histories" is a defensible position. "We provide personalised support" is not, because that's exactly what the AI tools will claim to do.

Second, look at what your staff actually spend their time on versus what creates the most value. If your advisers spend 40% of their time on CV formatting and job searching, and DWP is about to procure tools that do both, those hours need to be redirected. The value your charity adds is in the wrap-around support, the relationship, the cultural translation, and the ability to work with people who fall through digital cracks.

Third, think about which tech partners could help you deliver your specialism more efficiently. DWP isn't going to build bespoke tools for every hard-to-reach cohort. But if your charity can demonstrate that it uses digital tools to handle the scalable parts of its service while its staff focus on the high-trust, high-complexity work that AI cannot replicate, you have a much stronger case for continued funding.

Fourth, register for the 2 April briefing if you haven't already. The deadline is 11am on 31 March. Even if you're not a tech supplier, understanding what DWP is buying tells you what it plans to stop paying others to do by hand.

The JCS procurement isn't a threat to charities that know what they're for. It's a threat to charities that have been doing general-purpose employability support without a clear specialism. If your charity works with people who face barriers that a digital tool can't solve, language, trauma, cultural adjustment, complex health conditions, the next six months are a chance to make that case clearly, while DWP is still deciding what the service looks like.