IBM CEO: AI will help government do more with less without compromising quality

 

Q: What are the main opportunities for AI in enhancing citizen experience, and how can IBM support government initiatives?

NH: First, we need to understand what citizens value and what they find challenging about the current way in which public services are delivered and think carefully about how we use technologies like AI to do away with those frustrations and deliver the best of what is valuable.

IBM automated workflows for the U.S. Department of Veterans Affairs, cutting claims processing times by 90%. This not only sped up benefit payments but also freed staff from manual tasks, allowing them to focus on improving citizen services.

We see similar impacts in the UK. One of the case studies of which we are proud, is the work at Universities Hospital Coventry and Warwickshire NHS Trust.  The project focused on identifying and implementing improvements in the outpatient journey, from referral to discharge. AI allowed us to identify simple changes which could make a significant difference. Adjusting the timing of SMS appointment reminders reduced missed appointments and allowed the Trust to release 700 additional appointments per week over three months post-implementation. Imagine what would happen to NHS waiting lists if we could release this scale of additional capacity across the whole of the NHS.

Q: How can AI help achieve cost savings without compromising service quality?

NH: AI is a powerful tool for balancing cost efficiency with quality. Back to the example from the NHS, not only did we see service improvement but across the project, the team identified over 40 insights and interventions with the potential to save £2.8 million annually. The use of a patient-facing virtual assistant reduced calls to the patient access team, and we anticipate that alone could result in cost savings of £175,000.

AI can significantly reduce costs in departments by streamlining behind-the-scenes activities, like experiences in the private sector. At IBM, we have focused on enhancing its internal processes, particularly in HR, by implementing AI tools. The AskHR chatbot, based on the watsonx AI platform, has improved HR service delivery, and provided employees with easy access to HR services. This initiative has saved IBM’s HR department 12,000 hours over the past 18 months, with significant improvements in satisfaction rates. Back-office savings across HR, finance, operations, and IT are being delivered across all industries using AI

Q: In more detail, how can AI solutions help reduce cost drivers such as failure demand?

NH: Many of us have experienced the frustration of having to resubmit incorrect forms or needing to call a helpline to try and resolve an issue we simply cannot work out how to solve online. The reason this is a particularly acute problem for the public sector is that it often means your request must go back to the entry point of the process wasting all the time and effort that went into processing up until that point and adding to backlogs. It can also mean us submitting data, for example, for passport applications or driving licenses, when the Government already has these details.

There are many ways that AI can help in these scenarios. I have already talked about digital assistants being used inside IBM. But we have plenty of other examples from the private sector where AI-powered chat bots in banking and other sectors are helping to answer queries quickly and accurately. Our own AskHR system handles more than 94% of HR queries, resolving more the 10m interactions a year.

AI can also help identify shared areas where customers themselves might leave out key information or make simple errors. This can help process designers rethink forms to offer greater clarity and for guidance to be improved to reduce customer-initiated errors before a form is even submitted.

Q: The concept of "Total Flow" has been proposed in public service delivery to reduce waiting times. Could you elaborate on this approach and the potential role of AI in improving process efficiency?

NH: The capability of AI in this instance is related to its ability to see patterns and connections within complex data sets that would not otherwise be available. Enabling government departments to share data opens many new possibilities to verify citizens, join up experiences and deliver more personalised services. This is one area where I think AI will be increasingly transformational in the public sector as it will shine a light on how systems work as a whole. That means government will be able to make more targeted interventions at the point at which there will be most impact.

Q: Fragmentation in data sharing between departments has long been a challenge. What can government do to enable more integrated data systems to support collaboration across the public sector?

NH: We were excited to hear this government commit to the National Data Library. The value of data has been understood for a long time but with AI there are new and interesting possibilities.

First, we should not forget that a lot of data assets in the public sector are not digitised and the ability of AI to turn paper-based data assets into digital formats quickly and cost-effectively will be transformational just in the accessibility of the insight that will be available.

Second, it offers the opportunity to avoid the time consuming, difficult, and expensive process of moving all your data to an individual location with the support of integration. We see AI as a tool you should be able to take to your data, wherever it is and that opens the prospect of integration that is “virtual” which itself should make accessing data simpler while also ensuring there’s proper controls over access and security.

Q: Modernising legacy systems presents both risks and opportunities. What strategies would you recommend for balancing the risks of modernisation with the need for efficiency?

NH: This really gets to the heart of why technology is a strategic issue for all organisations. Legacy systems are not just an issue in the public sector. However, I think our private sector clients have been quicker to understand that advanced capabilities are harder to implement in systems that are already a spaghetti of old code and with apps in various locations.

DWP is a notable exception. The departmental leadership realised that its technological transformation would need start with a holistic rethink of its IT estate, much of which was with third-party providers that had been supporting its systems for several years.

My colleagues at IBM helped DWP move more than 60 applications onto 1,500 virtual servers. This has allowed the team to be more agile and to drive up productivity, but their approach has also had much wider benefits. DWP estimates that since the start of the initiative, it has reduced the number of major incidents by 73%. This has reduced user hours lost and has already saved more than one million user hours for the department.

One of the things we do not often speak enough about is the risks and vulnerabilities that are created in legacy systems. Technically that requires organisations to continue to patch and update their systems to keep them secure. This might seem like an unnecessary cost, but we know from our own studies (link to IBV) that the cost to organisations that are not paying enough attention to their security can be extreme.

Ultimately, we need to find a balance to ensure we are protecting the organisation while building the foundations of infrastructure that will release new opportunity and deliver the kind of value that I have already described.

Q: Ethical considerations are essential in AI deployment within public services. From your perspective, what ethical issues should government officials keep in mind?

NH: It is important that governments think carefully about the governance of AI across the public sector. It is something we think very carefully about and have done for several years. Our Ethics Board is a critical mechanism to hold our company and the team at IBM accountable for the ethical development and deployment of technology

It is also one of the reasons we have been transparent about the data that goes in to our own AI model and why we have developed other tools to allow companies and governments to manage issues such as fairness and bias and monitor whether the AI is performing as expected. We think the demand for this kind of transparency and oversight is only likely to increase and is critical in the public sector.

Q: How can government departments ensure transparency and public trust in AI-driven decisions?

NH: This is at the heart of the challenge of AI adoption. If people do not trust this technology, they will not want to use it and we will forgo all the benefits that are on offer, whether that is cost savings for service improvement. Transparency is an important part of how we build that trust and confidence.

Any discussion about AI should start with data and that means the conversation about transparency should start there too. Understanding what data you have, what is in it and what might be missing are fundamental. Bringing that data together and preparing it properly improves the relevance and precision of your AI output, but it also allows you to be clear about what is being used by your AI model.

Transparency, as a concept, is something that we have built into our AI capabilities, including our watsonx platform. That means we can be confident we don’t have hate speech in the data we use to train models, it means we can show our clients exactly where an AI-generated response came from, and it means giving our client the capabilities and tools to see, in real time, whether their AI is performing as they expected.

Q: The public sector faces a growing need for technical skills. What steps should departments take to build and retain this expertise internally?

NH: There is a shortage of specific digital skills across the economy. That said, I think one of the biggest challenges is how we upskill the existing workforce to adapt to the changing technological landscape. Partly that is a question of understanding the capabilities of these technologies. You do not need a PhD in computer science to understand what AI can do. It is also partly the approaches that we are taking to hiring, training and skills development across the whole of the economy.

These issues are going to be increasingly critical as government and industry are going to have to work together, in much closer partnership, to use this technology to best effect. While we bring a certain type of expertise, the public sector has its own expertise too. We are going to have to find ways to share a common set of goals and a shared language to make the most of what each brings to the party.

Q: Looking ahead, what are your aspirations for AI's role in transforming public services, and how do you envision IBM contributing to this journey over the next five years?

NH: IBM has been a proud investor in the UK for over 100 years. In that time, we have seen so many technological changes. Throughout each of these waves our aspiration has really been about unleashing the potential of science and engineering to solve some of the most intractable problems in the economy and society.

That aspiration remains the bedrock of the company, and I think in this we share similar values with the amazing people in the public sector who are making a difference every day. So, more than anything, I am looking forward to what we can do, in partnership.