Harsh Browns

The South Bank

Last week was my first full week working simultaneously for two clients – a step forward for my small consultancy¹ biz.

Rapid discovery

With Sport England, I’m looking at bottlenecks we might solve by applying a new data strategy that has been crafted. The goal is to think big, start small, learn fast and scale what works (as you might expect), showing how better use of data and digital ways of working can not only have immediate impact but help solve problems sustainably.

The week was filled with speaking to people, digging into their processes, understanding their challenges and what they’re working with, plus identifying opportunities for easing bottlenecks. Though we only spoke to three teams, everyone’s dedication to their work has been impressive, and you can tell the ethos behind the organisation is driving everyone’s work. That public spirit is always a joy to find, and it’s infectious too.

The main blocker we’re facing is access to people, source data and systems, which is more than a little prohibitive. But one team has shared some documents, access is being sorted by IT, and we’re looking at other ways to improve flow. It’s tough to chase people when everyone is so busy, but we really want to be effective, so getting hold of material to work with is key.

There’s plenty we can prototype and play with once we’ve got that, and I was impressed that Sport England had a mature AI policy and guidance for building in line with the AI playbook for government. What’s more, teams are happy to be supported by AI but believe decisions should be made by humans. AI literacy seems to be high, and one team are experts in prompt engineering and chunking data for AI analysis – despite not having any background in IT or digital. Pretty good!

Iterating the handbook

The majority of my time with planning.data.gov.uk was spent iterating the service handbook.

The bulk of the work is continuing to translate the value stream mapping we did into a high-level overview, calling out the stages of production on the platform and how each team contributes, including some of the individual processes at each stage.

It’s a bulky piece of work. Not only am I trying to document the flow and processes we’ve established over the last two years, I’m also trying to make it understandable to anyone on any team – so they can see their work in it and how it contributes to achieving the platform’s vision.

On Friday I added the team structure, creating an onboarding planner for new-joiners, and making the handbook more usable on mobiles plus adding search.

Leaning on GitHub Copilot agents too. Though I could have done a couple of those technical tasks myself, I’ve only got a few weeks left and speed is of the essence. Getting help from the bots is necessary.

Co-working

We’re one week into the experiment of using a co-working space instead of working from home, to switch up the energy. It’s too easy to get distracted at home and, to be frank, I’m in relax mode at home. Motivation is a lot harder.

Slapping cash down on a co-working space has created a good incentive to get out of the house for work. The place I’ve chosen is a good space to work in too: good desks, ideal desk lighting, daylight, coffee, snacks, and more. It has four types of working space: desk, library, lounge and phonebooth. Variety!

The background noise has been helpful. Plus other people working in the same space. Been craving the proximity of other humans recently. Though I’m not in the same space as the people I’m working with, being around other people working has been an OK proxy. (Looking forward to getting back to collaboration in close quarters soon.)

I’ve been way more productive, it’s been a good investment. On Thursday and Friday I couldn’t go in to the co-working space and felt lazy again. Took myself to a coffee shop near home on Friday and instantly became productive. Evidently I’m going to perform better by separating home and work environments. Mad to think I worked remotely for two years!

Is this the slow degradation of hybrid working on motivation and morale?

Running

Last week I was nursing an ankle injury all week, meaning I missed a couple of training sessions to give it some rest.

But then I had a hilly half-marathon – the Leith Hill half – and although I’d intended to take it easy, the course was so much fun, I just went for it!

10km up the hill, 10km back down, with a rounding of a smaller hill for good measure. We ran past beautiful Surrey cottages and streams, up through misty woodlands and along gravel tracks. The descents were tons of fun: I was flying down those sections, happy with my form and foot placement.

Though my ankle was a bit tight this morning, it has warmed up over the day and is feeling good. I’ve got an osteopathy appointment to get it checked properly, but it feels like it shouldn’t be a problem for the marathon in 6 weeks’ time.

Journal prompts

Bought a pack of journal prompt cards for fun. Going to pick a card at random and see what comes out.

OK, that didn’t work! Let’s try again.

The sorrows and problems I thought I would never overcome – but then did

Bit deep that. I thought I’d never overcome my hip problems from 2024, but I have through conditioning, focusing on form, and trusting the process. It’s much better now.

Maybe that’s all you need. Discipline, practice, and trust in the process of making things better.

Bookmarks

• Beyond the vibes, 5 mins. Think piece from Ben on how AI-assisted tools may change how public services are built by making it faster to turn ideas into working software. However, speed alone is not enough; teams must maintain strong standards and trust to ensure services are safe and reliable. The real challenge is learning quickly, working openly, and earning the confidence to move faster without losing quality – something many departments could have done better with over the last few years.
• The Demand Machine: The Realities of AI-Powered Public Service, 11 mins. Excellent piece showing how is likely to create higher demand on public services, which can overwhelm public agencies if they do not plan well. Governments need to prepare by adding staff, budgeting more, and keeping humans involved in decisions – not using AI as a cost-cutting tool.
• Samsung’s End-Of-Cycle Laundry Machine Song Has A Surprising Backstory, 3 mins. I have a weirdly positive relationship with my washing machine and tumble dryer, thanks to this little tune it plays – which also tells me when it needs cleaning out. Strong recommend.
• The (post‑it) notes you don’t play, 3 mins

Footnotes