Patryk Kosmider -

GDS looks at potential for using voice technology on website

The government has begun using machine learning to build a taxonomy hierarchy on and is looking at the possibility of deploying voice interactions

The Government Digital Service (GDS) is working on future-proofing the government’s public service website though measures including machine learning and voice apps.

The website, which went live in 2012, is often hailed as one of GDS’s big success stories, having resulted in huge efficiency gains and making it easier for citizens to access government content.

During a press briefing at the government’s Sprint 18 conference, GDS’s head of, Neil Williams, said work is ongoing to organise and simplify information on the website.

One of the ways in which the government is doing that is by building a taxonomy hierarchy, so that when a user lands on a page, it is easy to navigate around and look at related content. The work first began with the Department for Education and its agencies to group and organise their content manually. But it proved to be too big a job.

“We have been working on that, initially in a manual way, which involved a lot of engagement across multiple organisations simultaneously to get them to look at content together and group it and describe it,” said Williams. But it was soon discovered that this would take too long, he said, and that getting people together in a room to sort it out jointly was “nearly impossible to make happen”.

Instead, six months ago, GDS decided to use  machine learning to accelerate the work, creating an algorithm that helps to tag the content and develop a taxonomy.

“We are now at the point where we have got a taxonomy of everything, and all content on is now tagged by what it’s about,” said Williams. “In some areas, that’s good enough to be used, and in some areas it’s for development, but in the coming months we are going to get all that picked up so the whole cycle is powered by a single taxonomy.”

It’s about organising and simplifying information, he said. “By doing that, we’re making sure we are future-proofing government service delivery all the way.”

The government is also looking at the potential for third-party apps and websites, as well as voice technologies.

Read more about government IT

“Increasingly, it’s not really going to be important that people come to the website and browse around,” said Williams. “We expect people to be meeting their needs through third-party apps and websites receiving our content through APIs [application programming interfaces], but also new technology platforms and voice interactions [such as Amazon Alexa and Apple’s Siri].”

The work on voice is still in the discovery phase, he said, but GDS hopes to know more about what it needs to do and how to do it by the end of the quarter.

“The use cases at the moment aren’t terribly clear as to what are people are looking for by voice, specifically for government services,” he said. 

“We could go and build Alexa or Cortana skills, but we want to look at this at scale for the whole of government.” 

GDS is also looking at popular services, as well as those affected by Brexit and and structuring those as step-by-step, structured services.

“There are about 400 services we have identified that we want to apply the step-by-step navigation plan to,” said Williams. 

He described it as a “highly collaborative effort”, bringing together the organisations involved in that particular service, many of which have never spoken to each other before.

Departments are also being empowered to manage their own content, he said. “We are providing them with the tooling they need to understand, using data insights to prioritise, and continually iterate the user journey and reacting to what is changing.”

Read more on IT for government and public sector

Data Center
Data Management