Digital Centre of Government: first reflections

I’ve been appointed to an external advisory panel to support the design of the “digital centre” in DSIT (currently a smooched together combination of GDS, CDDO and i.AI). I put out a call on Bluesky for reckons and pointers that has had quite the response, summarised here by Tim Paul but you should go read all the responses. I want to try here to distil some of the topics, questions and opinions around the design of public services, technology support, and what DSIT needs to do as the “digital centre”.

By way of caveat, these are my current reckons. They are not necessarily what DSIT thinks, nor will think, nor will do, and I’m very aware I don’t have all the context or experience so may be wrong or change my mind as I learn more. I’m getting this down now to prompt further discussion and help me refine what to advocate for.

At a high level, it seems that the things that need to be figured out are:

1. How does this Government want to approach public services?
2. Where and how can technology best support this approach?
3. How can DSIT, as the “digital centre”, help the public sector use technology well?

Then there’s the meta point about how DSIT should engage with broader communities and stakeholders to figure out the answers to these questions. As Martha said, I’m certain the answer to this is “in the open”, properly engaging with all the expertise, experience and thought that’s been building up over the last decade and more, both in the UK and elsewhere, within the “digital public services” community and outside it.

I am strongly of the opinion that the answers to all the above questions are not things that can be decided entirely up front, in a short time period, nor by one team in one department. These are shared understandings and visions that have and will develop through discussion, debate, experimentation, and conscious reflection and learning. This is something that has to be done together.

So I’m heartened that the DSIT team is taking maximising engagement seriously. And by the energy outside that team (as evidenced by responses to that Bluesky post). I’m also realistic that it is super hard to remain open to a barrage of sometimes challenging, critical, contradictory and even emotionally charged opinions. Hard to hear the people who aren’t part of the conversation. Really hard to share work in progress and unformed thoughts, particularly while also dealing with potential media and political attention. I think that one of the things we can do as external advisory panel members is take on some of those risks and challenges, because we are not civil servants, operate in slightly different spaces, and are not ultimately making the decisions, so have a bit more freedom in how we operate.

With that context, caveats and expectation-setting in place, on with some of my thoughts.

How does this Government want to approach public services?

I’m phrasing the question this way because:

1. Public services cannot be entirely digital
2. The way public services work is a big-P-Political choice

Fundamentally, we need service transformation, not just digital transformation. I oscillate between thinking the “digital centre” should lead on that because digital teams seem to be where most service design is being done, to thinking it’s a wild overreach and DSIT should stick to its lane while advocating for the broader service transformation that’s required. Perhaps this is a point for minister-level negotiation of the “digital centre”’s role. I’d love to hear more thoughts about this.

The Political questions about service design include:

• Which types of public services does this Government want to prioritise? Ones used by businesses (for economic growth)? Small businesses? Working people? Underserved people? Ones that help advance the Missions? Ones that have the highest volume?
• How does this Government want to treat the public? Autocratically? In partnership with them? In service to them? As if the public is basically honest or basically fraudulent? As customers or as citizens? Are we aiming for something relational or transactional?
• How does this Government want to deliver public services? By the public sector? By the private sector? By social enterprises and charities? Through single, controllable, routes or multiple, less controllable, routes? Prioritising quantity or quality? Directly or through intermediaries?

It might be that there are different answers to these questions in different contexts, but I think it would be really helpful if we knew what general style this Government wants to adopt. That’s not entirely clear yet, at least from where I’m standing.

There are answers I would personally like and argue for. For example:

• Prioritise based on Missions, implying that there should be a strong tech-aware presence in every mission board / delivery team.
• Aim for relational, humanising services in which people are treated as citizens (not consumers), with respect and dignity, with government in service to and held accountable by them.
• Meet people where they are by delivering services through multiple routes, favouring social enterprises and small businesses and carbon neutrality, and recognising the important role of intermediaries and supporters that help people and businesses access public services (about which more below).

That’s the perspective I’m coming from as I’m thinking about what’s needed in technology in public services, but I’m really not sure it’s how this Government is thinking about things at the political level.

Where and how can technology best support this approach?

I’m not going to cover everything here. I’m not an expert in end-user-facing digital service design; many of the details are covered in Richard’s book and live in existing guidance and in the heads and practices of people making these kinds of services work day to day.

But I do want to expand our thinking about who we should be supporting with this tech transformation.

I think we can unconsciously default to talking about members of the public doing life admin online (driving licences, tax returns and do on). Perhaps we also think about organisations like businesses and charities, especially given the emphasis this Government has on productivity and economic growth.

A few people have rightly highlighted tech supporting public sector workers themselves. A lot of delivering better public services is helping the people doing that delivery to do it better and, to get through all the current backlogs, more efficiently. This with important caveats about developing more relational government for citizens and ensuring public sector jobs are satisfying “good work” rather than taking a gung-ho “do everything faster with AI” approach. This support includes custom tools for backend public sector tasks (eg consultations, grant-giving, registration) and making it easy to adopt and use good off-the-shelf / no-code products and services.

Also, we tend to focus on people directly using digital services provided by the government, on their own behalf. I think (as above) we should be paying a bit more attention to intermediaries. Some people who don’t use digital services will shift to using them if we improve broadband connectivity, build digital literacy, make services more accessible and so on. Others might use offline versions of services, such as paper forms. But for many, and particularly those who most need to access public services, the best kind of support they can have is a real person to help them.

These intermediaries might be slightly-more-digitally-savvy carers. They might work informally in communities, eg GP receptionists, librarians or charitable community support workers. They might have a more formal support role such as Citizens Advice advisors or call centre workers in the public sector. They might even (I imagine especially for businesses and charities) get paid by people to help them, such as accountants or tax advisors. Regardless, finding ways to help these intermediaries help other people, without allowing them to overstep or see more of someone’s lives than they need to, feels important to me.

Somewhat related, I think we should be considering the provision of technology (particularly data, but also digital services and maybe things like compute) as a public service itself. We should be looking at how the provision of data and digital services can support the creation of software by developers in the private and third sectors as well as other parts of the public sector. These could be tools that provide other interfaces onto government services (for example, look at accountancy software that automatically submits tax returns). It could also be B2C and B2B products and services, particularly where they’re needed to advance Government’s Missions.

Finally, we should be building in accountability and continuous improvement to these systems. We need to keep in mind that technology is imperfect, as are our approaches to public services. Building in things like monitoring, handling of exceptions, supporting independent research and investigation, enabling complaints and appeals, and so on, are all part of what helps public services get better over time. The users here might include auditors, parliamentarians, researchers, investigative reporters, campaigners, lawyers etc.

Understanding these different types of users is important because they’re the people to talk to about the types of digital services that would make the most difference to them. I’d like to see these different groups engaged in co-designing the vision for digital public services, not just as subjects of user research for individual products.

How can DSIT, as the “digital centre”, help the public sector use technology well?

This is the big question. There are loads of opinions out there about the biggest barriers to people building great digital services in and for government. I’ll note it’s often easier to identify those challenges than to work out what the digital centre can reasonably do about them. Many challenges aren’t new, and it’s important to understand why previous attempts didn’t work as well as they needed to, or got rolled back over time; whether circumstances are any different now; and from that understanding, what approaches to try next.

But the first thing I want to talk about is attitude, in particular to the rest of the public sector. We can debate the extent to which this was true back in 2010ish, but it’s certainly the case now that there is a huge amount of good digital/data capability across the public sector and beyond, and that policy and operations people know their shit too. I believe that to get the most out of the system, that existing capability needs to be recognised and honoured, and any work at the “digital centre” needs to be facilitative and empowering.

I think demonstrating humility and generosity will help the centre be more effective, particularly at reaching across traditional fault lines, such as into devolved administrations, local government, arms length bodies, the NHS, the big digital departments DWP and HMRC, and into non-digital professions, for public-sector-wide transformation. Think more like community and movement building than top down waving of big sticks.

With that in mind, from the range of suggestions and blog posts I’ve read, plus my own reckons, it looks like there are three types of activity that would be helpful:

1. Enabling community and learning: This ranges from supporting formal learning through a revamped Digital Academy, through to enabling peer learning with truly cross public sector communities of practice. I would include in here more public dissemination and showcasing of successes and failures (bright spots and bear traps) through comms and events. As well as learning for digital professionals, it should include programmes for leaders. Inserting things into existing training might give greater reach into non-digital areas, and it might be worth thinking about bringing non-digital topics into training for digital professionals. Secondments into and out of the centre, and across into different parts of the public sector, might be another approach, particularly if it would give more space for people to experiment and take ideas back into their home organisation.
2. Stewarding the commons: It would save everyone time and effort if we could reduce duplication across the public sector. Here we’re talking about research, patterns, checklists, standards, libraries, platforms and so on. This isn’t (necessarily) the centre doing things for the rest of the public sector or heavy-handedly telling them what to do, it’s helping by coordinating activity to jointly develop (or procure) the things that everyone can benefit from. There are examples of this, like Notify and the Design System; there should be more, and with a more distributed/shared ownership model (in my opinion). I’m particularly interested in how the demand and supply for these should emerge and become formalised (see this great list from Frankie Roberto). There’s also a bit here about enforcing standards – ones close to my heart would be around accessibility, risk assessment and participation – which needs to happen but also needs to be done constructively.
3. Breaking down barriers: There are a bunch of challenges around tech adoption, where the digital centre needs to have the political gravity and persuasive power to successfully argue for changes that make things easier and better. The ones I see mentioned most often are:
   1. Improved digital procurement, to keep pressure away from procuring big system integrators. (I think the new Procurement Act, and this Government’s focus on social value provide near-term opportunities here, including to be able to properly factor in the carbon impact of generative AI etc.)
   2. Business case / budgeting that fits better with tech, such as funding teams not projects; covering maintenance, not just development; and giving room for innovation and experimentation. Changes here are also vital to unlock the kind of cross-government working, as above, that prevents duplication in this space. And to ensure we fix the foundations, don’t just focus on the shiny stuff.
   3. Putting in place a whole-service transformation approach that means digital teams aren’t left just putting lipstick on pigs. I’m somewhat hopeful that the cross-government and multidisciplinary ownership problems around this might be tackled (to the extent they can be) by government getting good at Mission delivery. But that’s pretty central-government focused and it’s likely more will be needed.
   4. Unlocking comms to enable working in the open, because making things open makes them better, and it’s a real enabler of both learning and cross-public-sector working.

What’s next?

I outlined at the top that the timeline for this “digital centre” design activity is pretty rapid. I can’t make promises on behalf of the DSIT team, but I have been and will be pointing them at all your thoughts and doing my best to make sure they’re incorporated into the plans.

I’m also hoping that the DSIT team will be able to get something more official and semi-structured in place for engaging with all the expertise that’s out there more directly, as well as with people outside the rarified climes of Bluesky.

In the meantime, do please tell me where you disagree or I’ve missed things in this post, any other things you think of, and where there’s great things happening that should be more widely adopted. My big questions at the moment are:

• How to (from DSIT / this project) make space for whole system redesign where that’s needed.
• How to manage the budgeting side of cross-government platforms, data provision etc. in ways that support shared ownership.
• How to balance a drive for standards compliance with a constructive, non-patronising attitude.
• How to give the right kind of space for innovation at all levels of the system.

More importantly, talk about this with your colleagues. What vision do you have for how public services should work? Who needs support and how could technology practically provide it? What support do you need to make that happen? This moment is an opportunity to take a breath, reorient, and shape the future.