Image source: Swindon Council
Interview: Head of emerging technology Sarah Peña talks about how it approaches projects and the value of horizon scanning
There is always talk about the need to look at the business problem before the technology for digital projects, but Sarah Peña says her experience at Swindon Borough Council has shown that looking at the tech first can often bring good results.
As head of emerging technology and business improvement she has overseen a series of projects over the past four years, some of which have been prompted by the team highlighting the potential of a technology as the first step.
“We’ve had success both ways,” she says. “Sometimes with a service area they have so much in the day-to-day delivery that they don’t necessarily have that awareness of what might be out there to help them.
“Having a technology viewpoint – although it is still always the enabler – it can be that technology is a facilitator to look at the entire piece. All the usual stuff in terms of the business engagement, commitment to do something differently, and the processes ultimately have to be there; but it can start from the tech.”
Peña – who has previously worked in central government and the private sector – set up Swindon’s emerging technology programme in October 2019, based on her appetite for looking at new technologies and with the chief digital officer at the time selling a proposal to council members. She obtained seed funding of £100,000 and an agreement to have 12 months in which to show it could be worthwhile.
Doing it at pace
“The agreement was that I would be given the funding and able to create a model of working, on a blank piece of paper, and would be supported to work differently,” she says. “The chief executive and CDO gave me full support to work with autonomy and differently so we could do stuff at pace.”
For the first few months she worked alone, was then joined by one of the council’s geographic information systems team, and since then has taken on a business analyst and low code developer. These were both internal moves, reflecting the desire to give people the chance to advance their careers inside the council.
This was accompanied by the early emphasis on coming up with the ideas for projects.
“For the first year we generated our own work, sourced technology we thought would benefit the council, then engaged individuals who we thought could generate use cases and worked with them.
“Around month six it generated some good momentum, trust had built, some internal knowledge was growing around what we could do, and we now do about a 50-50 split. We still do things self-generated and we’re still horizon scanning to look at opportunities, and we have 50% from the business around problems it faces.”
Working model
The team has adopted a seven-step proof of concept model, with a blended methodology of agile and lean and an emphasis on working closely with service teams.
“We like to spend time on the ground with them to understand all the problems and what a day in the life of what they do looks like. That time can vary, may only be a few hours, but there is a structure behind it.
“In the big majority of cases, around 95%, we go from conversation to something being trialled in a few months.”
A project typically involves a three-month trial, which can vary depending on the time needed, followed by the production of a report in the next two weeks. This includes an outline of the original aims, the agreed key measures, results of the works and learnings, and an assessment of what a ‘go live’ product would look like.
Early projects
The conversations on projects began around highways and waste, and the first proof of concept prompted the need for a rethink. It was aimed at harnessing computer vision AI for the proactive identification of potholes in the roads, but the team discovered the technology could not provide the data needed in areas such as depth management.
But they quickly saw it could be used in analysing images of fly tipping, along with a reporting function with a location feature on the council website, to prioritise the workload of clean-up teams. By 2021 this had produced significant savings in staff time and fuel costs, and won widespread approval from people using the service.
This was followed by the development of a bot to help manage the spike in applications for free school meals during the Covid-19 pandemic. The work had already begun when the pandemic broke out, the team had to accelerate the work, carried out the testing over the weekend and managed to launch the bot on the following Monday.
There have now been 25 projects, of which two have stood out for Peña: the development of a machine translation service application to support multicultural communities in the borough, and the use of drones for the inspection of the roofing of public housing – both of which have been previously reported on UKAuthority.
“I really like drones for roofing inspections,” she says “It’s really transformed that role for them. I love the fact that it opens up future employment opportunities for people who wouldn’t have been able to fill that role beforehand.
“I like how the communities have really engaged around it. It has brought benefits to the council but also been fantastic for workforce benefits, resident engagement and support.”
Easy Read
She also points to a project with AWS, due to be announced soon after the conversation, to develop an Easy Read AI solution which converts complex documents into an accessible format. It has been carried out with a local volunteer focus group, whose members have a range of learning disabilities.
The tool can be converted into 75 languages and there are hopes it can also help people with low level literacy or cognitive impairments.
Only one project, for adaptive lighting of traffic, has not been taken to the live stage. It was put aside when it was found that the council’s LED lighting programme was already providing benefits in line with those expected.
Another, for sentiment analysis of local public perspectives during the pandemic, was regarded as successful; but ideas of further deployments were dropped when the business case was felt not to be strong enough. Peña qualifies this by saying there could be scope for it to be brought up again.
It is also significant that the projects are carried out with the intention of making the solutions available to other public sector bodies on an open source, licence free basis. Peña says the machine translation solution, available on GitHub, has been picked up by a number of local authorities and organisations in France, the Netherlands, Poland and the United Arab Emirates. In addition, a couple of other councils have used the fly tipping solution.
Looking forward, Peña sees a lot of potential in generative AI – “There’s going to be so much more to do in that space and we’re just touching the beginnings of that one” – and says the team is also working on an AI initiative around information governance, and a project on tech enabled care.
Key lessons
In conclusion, she says the experience of the past four years has provided three key lessons for local authorities looking to build solutions on emerging technologies.
“I would say pace is important as it helps you garner the support quite quickly. Try to get a couple of good early wins. It’s one of the reasons I put things like Translate out there. If there’s another local authority looking to do something like we’ve done there is stuff there that could give them a head start.
“Another is measuring benefits. I’m able to talk clearly about benefits for each piece of work and overall. It’s hugely important that you’re able to clearly articulate the impact the work is having.
“Also, there’s building ecosystems. One of the things I take time to do is build an ecosystem, creating relationships with others to have that collective wisdom. It’s inside the organisation and externally. It all helps support what you are doing.”
