A conversation with Sigrit Siht, Director of AI & Data at Digital Nation

Edition #113 Digital Nation's Director of AI and Data talks about the role of the private sector in digital government ecosystems, and seizing the low-hanging AI fruit...

Sigrit Siht is the first representative of digital government’s private sector that interweave has sat down with in our interview series. Digital Nation, the “digital change advisory network” where she is Director of AI and Data, works across four continents supporting governments in their digital journeys.

However, while Sigrit is working outside of the government today, it would be hard to accuse her and her colleagues of not having been there and done it themselves. All of the technical consultants currently working for Digital Nation – under the leadership of Siim Sikkut, former CIO of Estonia – have previously worked for the Estonian government, at least four of them in the heart of e-Estonia’s digital transformation teams.

We begin by talking about the role of private sector actors in the digital government world. As advisors, Sigrit says, it is important to “understand that although you have a technical knowledge base, there are lots of rules in the public sector that only civil servants understand”. It is important to be humble – to focus on co-creation and to have “conversations built around a shared understanding”.

This is especially true when it comes to talking about best practices in the context of Estonia. “People can sometimes see Estonia as a happiness world, where everything is perfect”, she says, “but it is important to help governments understand that the goal shouldn’t be to be Estonia 2.0 or even 1.0”.

Instead, the goal is to help governments understand their own cultural differences, and reckon with their specific pain points for themselves. “Best practices serve the purpose of getting people thinking, but that’s where it should stop […] you should set your own requirements and your vision within a closed system”.

In this environment, external expertise can bring a different point of view. Sigrit talks about her experiences of working on Burokratt – Estonia’s interoperable network of public sector chatbots. Her team at the Ministry helped complete an assessment of 1,000 Estonians alongside a private sector partner, to help the government’s central digital team understand how different users (including other agencies) felt about the tech. Unsurprisingly, when government colleagues felt they were being interviewed by experts without vested interests, they were more forthcoming with their answers.

Making digital innovations interpretable across government

Much of Sigrit’s work focuses beyond technical builds. We spend a lot of our time together talking about two often overlooked, but major, themes of digital government: capacity building and awareness-raising.

When it comes to capacity building, Sigrit is clear that “it is a very continuous thing that needs to happen – the work is never done”. And yet, “if a new government comes in and talks about money historically having been wasted, it is one of the first things that tends to go”. The risk is that this contributes to creating an understanding and communication gap between technical and non-technical teams.

Bridging this gap is very much a two-way process, Sigrit is keen to stress. It is just as important to make things understandable as it is to understand them. She talks about experience of her twin sister, a systems analyst at an IT company. “She would take the train to work and, as they got closer, it would be filled with technical people and programmers all talking their language”. But after meetings, non-technical experts would come up to her and thank her not for her knowledge but for “making things understandable, rather than talking in big words”.

After all, the end goal is “to first understand the topic yourself, and then make it accessible to everyone – nobody should want to gatekeep knowledge”.

In digital government terms, the stakes of capacity building in this way are high. A communication and understanding gap can lead to poor user design, low uptake, and diminished levels of trust in a service from the public. At worse, it can create a perception of opaque algorithms or black box thinking, removing the “human face” from government.

This is especially important when it comes to using AI – examples in the Netherlands and Australia indicate the consequences of when this goes wrong. But even at a smaller scale, a lack of clear communication can prevent an ecosystem from forming and success from scaling.

To scale AI – says Sigrit – “it’s important to understand what is currently working and what could be done better”. “Reuse is a big issue for IT departments right now”, with Estonian public sector source code even published on GitHub. But if nobody knows where to look, or how to interpret that code, then nobody will really use it.

Estonia publishes its source code on GitHub

Seizing the low-hanging AI fruit

Given our conversation so far, it makes sense that when I ask Sigrit about what most excites her in the world of AI, her answer focuses on simple innovations. She is “excited to see a civil servant working on social welfare finding Microsoft Co-Pilot: using it to read a lot of documents, summarize and then write a lot of documents and plan”.

“It is not necessarily about finding a new flashy thing”, she says, “we already have a lot of solutions”. Instead, governments ought to “set aside the fancy project for now and say to a user ‘what are your problems as a teacher or social worker, what takes the most time?’”.

In the contexts of Digital Nation and Estonia, one example of this mindset in action is an AI Accelerator. There, Sigrit and her team “bring in teams from all different ministries, have at one table a product owner, lawyers, and people that understand current processes and barriers to them, and a seasoned data science specialist to help find solutions”.

Digital government problems are solved by conversation, she says. “If there’s a blockage in a pipe, you can put as much water into it as you want, but without solving the blockage the water cannot reach its intended target”.

With this, we come full circle, back to the idea of how the outsider can support digital teams. This need not be an external consultant, but could also be a central digital team going into a ministry. AI can even do this too, as Digital Nation has exhibited. Supporting digital government, she says, is about recognizing when there are two civil servants with the workload of five people, and stepping in to help them take a step back.