The European Commission’s Joint Research Centre (JRC) is currently doing a piece of work to develop an omnichannel framework for public services across the EU. A week ago I joined a workshop alongside people from the OECD, Portugal, Sweden and Denmark.

It’s a topic close to my heart. Beyond my years learning from the best at GDS I was fortunate in shaping the OECD’s Framework for Service Design and Delivery that was developed during this research in Chile, summarised in this Going Digital Toolkit Note and further embedded into the OECD’s Good Practice Principles for Public Service Design and Delivery in the Digital Age. That work laid the groundwork for the OECD Recommendation on Human-Centred Public Administrative Services. I didn’t work directly on the Recommendation itself (it kicked off in earnest after I left) but those earlier pieces helped to shape the foundations it builds on. Bruno shared the definition from the Recommendation.

Omni-channel refers to the approach to managing service delivery channels in an integrated, interoperable way to enable users to access the service they want seamlessly and with consistent quality across channels (such as websites, physical offices, self-service kiosks, video-calls, call centres, etc.), as opposed to a ‘multi-channel’ approach that refers to the ability of the user to access services through different entry points, often operating independently of each other.



OECD Recommendation on Human Centred Public Administrative Services (2024)

Now, through my work with DWP on the jobs and careers service, it’s great to be revisiting those ideas afresh, and in concrete terms. Part of our vision is to design a service that is ‘digital where possible; human when needed’. As an organisation with a national footprint of Jobcentres and a wide range of services that can (and should) be delivered online, we’re asking: how can we bring the best of in-person services into people’s palms and pockets, while at the same time bringing the best of digital into those frontline conversations for both staff and our users?

From tailored CV advice online, to guidance offered face-to-face, to the potential of personalised prompts and nudges through an agentic interface in the future, our aim is to meet people where they are, in the mode that works for them.. And for that, we need to be focused and ambitious in pursuing an omni-channel approach.

All of which is to say: thoughts, I’ve got a few. I wanted to write up the post it notes I stuck on the Miro board – it’s not a formal report, just an open reflection. If you’d been in that call, here’s what you’d have heard me say – but what would you have added (or objected to)?

What even is omnichannel?

Some people consider omnichannel and multichannel to be synonyms. But they’re not. Multichannel is just having different points of entry: online, phone, or in-person. But these channels often act in isolation. Ever had to repeat yourself when switching from website to phone? That’s multichannel frustration in action.

Omnichannel promises something different. It should mean you can start online, pick up the phone to clarify something, and then walk into an office to finish, all without repeating your story. It’s seamless. It makes sense. It’s designed around you, not around organisational structures.

In an omnichannel setting users can start, continue and complete their journey across digital, physical, and human channels without friction.

Beyond life events: whole circumstances

The life events model of organising services around particular moments like births, job loss, or retirement is a popular one. It can be a handy frame for zooming out from an individual transaction and understanding the broader situation someone might be experiencing.

Moreover, everything that occasions an interaction between government and citizens is probably a ‘life event’ of some kind. Life events don’t go far enough.

I also think that while they’re helpful, especially in establishing a working baseline, they’re potentially a bit narrow. They treat life a little bit as though it happens neatly, one event at a time. But real life isn’t tidy. People live in complex, overlapping circumstances.

A person renewing their visa might also be navigating a high-risk pregnancy, dealing with insecure housing, and trying to secure SEN support for their child all while working shifts in a precarious job. Yes, that might be an extreme case but it is plausible. Each of those needs touches a different part of government, but for that person it is all one tangled knot of circumstances.

No service is an island

Designing around isolated events is definitely an improvement over not doing that but how much might we miss the whole context? Earlier this year I saw some slides from GDS that showed a three step progression from Life events at one end to Life circumstances at the other (there was a step in between that I can’t remember). I haven’t found a blog post or anything from GDS sharing that thinking but it’s wise and resonates with my disquiet over life events.

In our jobs and careers work we’re exploring how we might move away from working out from upstream segmentation toward a highly tailored approach that can respond to the needs of the “segment of one.” If we can collate enough insight about someone’s circumstances and pair it with an omnichannel model, then we can hopefully connect the dots across systems and channels so that the support we provide really does respond to the reality of people’s lives. That’s certainly the ambition.

What are the core characteristics of an omnichannel approach?

  • Seamlessness: the user journey flows smoothly across channels without disruption
  • Inclusivity: services work for people who can’t or don’t want to use digital
  • Common infrastructure: shared tools and platforms reduce duplication and fragmentation
  • Policy-operations-delivery-strategy alignment
  • Transparent, ethical, consent-based use of data.
  • Flexibility: services adapt to user needs and preference rather than enforcing a specific channel
  • Deep organisational integration: putting in the hard yards on joining up responsibilities, funding, leadership, collaboration, people and not just tech and data

What are the enablers and constraints?

Technical

  • Legacy systems block integration due to missing APIs, inaccessible data, outdated contracts or arcane programming languages
  • Inconsistent solutions for core capabilities (identity, payments, notifications, forms, hosting, etc) across organisations
  • Data interoperability is limited by tech, policy or trust. Consent models, APIs, registers, governance, etc may be lacking

Client-related

  • Users shouldn’t need to understand government structures to solve problems
  • The public expect government to be joined up far more than we are; they don’t understand incoherent and inconsistent journeys
  • Language, literacy and accessibility are barriers – we need to think about the level of language capability we write for (I’m reminded of this incident with a Home Office letter that we weren’t sure about, let alone its actual recipient)
  • Digital by default is unhelpful; digital might be the starting point but that should be in support of how you make it possible to support people in the ways that work for them (as per our ‘digital where possible; human when needed’)
  • ‘Segment of one’ personalisation is a great ambition but relies on trust, good data and functioning tools and capabilities

Organisational

  • Siloed ownership of different channels leads to fragmentation
  • New channels can proliferation without strategic coordination or commitment to a vision
  • Funding and staffing structures inhibit joined-up service teams
  • Leaders are increasingly expected to be digitally native but are still susceptible to hype than attuned to the underlying constraints
  • Multidisciplinary should not be narrowly defined as being about UCD professionals (design, developer, delivery, etc); collaboration means recognising digital transformation as a fundamental part of everyone’s job rather than solely a special specialism
  • Ownership of whole problems and end-to-end journeys needs to be shared across institutions

Legal

  • Legislation often encodes old service models, making iteration difficult because the law itself says paper and signatures must be used
  • Regulating ‘what good looks like’ is more about building a culture that fosters and encourages good practice than enforcing and controlling compliance
  • Procurement is hell
  • Data sharing is hell (not everywhere but definitely in the UK)

Data

  • Feedback and insight into performance are non-negotiables
  • Performance metrics can’t just stop at satisfaction
  • What are the composite measures that can work at the central or national level and which are built up from definitions of performance owned and managed by individual service teams as they improve and perfect the experience for their users?
  • How do you build up to a national picture from a singular service and locally meaningful insights?

I could have said a lot on this but we were running out of time by this point. If you want all of my data thoughts then this series of posts from a talk I gave in Azerbaijan is for you.

Other:

  • Ministers want announceable changes but the most important work to make things seamless generally means solving hidden and invisible things. You can’t really announce good quality delivery
  • E-government ≠ digital government. Digitised ≠ digital.
  • True digital services are intentionally designed for flexibility, inclusion, and iteration—not just ported online
  • Some suppliers are still extracting rather than enabling value

So what now?

The job of taking those thoughts and the thoughts of the others in the workshop as well as all the other events that have been taken place falls to the JRC team. I’m looking forward to what they come up with.

It was an energising couple of hours to talk about the challenge of achieving a genuinely omnichannel public service. Done right, omnichannel services don’t just deliver better transactions. They make it easy for people to engage with government, and in doing so, they restore some human dignity, ease, and trust. That’s the impact of digital government on citizen well-being.

Those were my thoughts – a little polished from Miro but still largely unvarnished and offered in good faith. If you’d been in the room with us, what would you have added? What would you challenge?