How to get started with Artificial Intelligence: A Business Leader’s guide

Advances in Data Science, Machine Learning and Artificial Intelligence (DS/ML/AI) are radically disrupting how organisations work. This is not something that will happen in the future, it’s happening now. Specifically, Generative AI – the technology that powers applications like ChatGPT – are ushering in a new era of AI possibilities. Its application within your organisation will reset expectations of what you think it is possible to achieve with AI and how far it can progress your future ambitions. 

Over the last few years organisations working to streamline their systems have focussed on digitising their paper-based processes to improve productivity. Now they find themselves limited by some of the same issues of the legacy service. So, what next? Lean analysis used to be one of the most important operational productivity tools, but today in the context of AI use case identification it is a game changer to driving productivity improvements from AI.

At Kainos, we worked with a large Public Sector organisation who were keen to explore the potential of AI and what it could do to release value for them. They wanted to use AI to improve productivity and free up capacity for more creative work which will help them grow their service offering.

The main use cases identified ways of reducing repetitive day to day tasks, improving the quality and accessibility of information and removing waste or rework from the system. As a result of implementing these recommendations they will not only significantly improve the quality of their services and user experience, they will create space for innovation which could transform their organisation. Let’s take a look at the 4 key considerations you need to keep in mind.

Firsly there need to be business value release requirements for AI return of investment (ROI), whether these are user needs (customers or internal users) or operational KPIs. To help define these:

• Are there clear business plans? Without this, you cannot define the high-level business value expected of AI?
• What are the operational barriers to realising the business plans? By understanding the current bottle necks, waste drivers, volume drivers, the features which don’t work well and missed user needs, we can identify effective use cases for AI solutions.
• Supercharge AI and the ROI potential; commission lean analysis of your operations to inform digital service and product design, of which AI will be a feature.

It’s clear that the rate of AI capability is growing, the days of a 5-year operational strategy are over. A high-level vision or an organisation’s purpose can just about survive that horizon but the operational enablers in that strategy, for those looking to benefit from AI capabilities, will look very different within 18 months.

To transition to an agile operational plan, CEOs should align their high-level vision with their digital roadmaps which will enable iterative business value release from digital and AI.

Each digital roadmap should have a prioritised list of high-level digital functionalities, or backlog: this is a list of digital functionalities required by the organisation. Each is prioritised for its ROI using business value identified through lean analysis and technical feasibility criteria. The most agile organisations iterate these criteria periodically, based on the organisation’s ever changing business priorities and emerging AI principles. For those organisations which require them, this mechanism lends itself well to developing business cases for investment to deliver the specific outcomes enabled by the roadmap.

Once you have identified the potential AI candidate use cases, each one is prioritised for ROI and wider impact risk dynamically, as sometimes the ROI and risk can vary depending on the results seen in digital development & testing.

This diagram shows how your strategy, business plans, digital strategy and AI Principles need to influence your digital programmes.

We recommend your digital teams deliver in a `dual track agile’ methodology. Where iterative design and delivery teams work synergistically. 

Many organisations have significantly invested in data. They may feel that because they have access to performance business intelligence (BI) that they have everything they need for AI. In our experience, that’s usually not the case.

Data maturity and AI maturity are ways of assessing and benchmarking how well equipped an organisation is to leverage data and AI effectively. There are many versions of maturity assessments, but they all cover the same general considerations:

Data Quality and Usability. AI thrives on data; therefore, the quality and availability of data are critical for success. A digital strategy should be in place to define whether your organisation has a robust and secure data infrastructure in place. It should consider the quantity, quality, and accessibility of your data. Are you collecting relevant data? Is it accurate and adequately labelled? The data you need for performance reporting is not the type of data that is required for technologies like Large Language Models (GPT) for example.

AI Enablement. Implementing AI requires a diverse set of skills, including data science, machine learning, and software engineering. Your strategy should evaluate your organisation's current capabilities and identify potential skills gaps. Hiring and retaining top AI talent can be challenging, so consider whether you have the necessary resources. If not, you could explore partnerships with AI-enabling service providers. Building a multidisciplinary team is crucial for successful AI implementation.

Implementing AI is not a plug-and-play process. It requires careful planning, testing, and iterative development. Digital leaders need to consider the potential challenges that may arise during implementation, such as integration with existing systems, scalability, and change management. Develop an implementation roadmap, allocate resources effectively, and set realistic expectations regarding timelines and outcomes.

The impact of AI will not only be felt by your organisation but also by the users of your products and services. A CEO should lead the development of and own the AI principles. They should engage collaboratively and iteratively with internal teams, customers, users and those whom your products and services impact. Getting buy-in is critical to a successful outcome.

The principles should cover themes including:

• The way in which you want to start using AI in the organisation e.g. decision augmentation – providing information more effectively to a human decision maker. Or you could consider a decision automation process that is mostly automated with human intervention by exception. Many organisations will choose to iteratively increase the automation of decisions following a period of validation of decisions (augmentation).
• What your strategic workforce plan will be when AI makes processes quicker: e.g. growth strategy: new revenue streams ready to go awaiting resource, or savings strategy: reduce costs and rationalise the operating model.
• Risk appetite: balancing quantity vs quality, assessing the effectiveness of AI vs current processes in a controlled environment will allow you to tailor your risk appetite.
• Ethical considerations: articulate how responsible AI decision making will be proactively embedded into the organisation including ttransparency of methods, accessibility and mitigation of bias into new data collection and processing. https://www.kainos.com/insights/blogs/implementing-ethical-principles-for-ai-in-defence2

In general, when principles that protect employment and focus on growth or ESG deliverables are put in place, engagement, buy-in and support will follow.

Indecision and lack of clarity in the purpose of developing AI use cases at an early stage, could drive the wrong behaviours and motivation. So, it’s essential for the leadership team, as a whole, to share clear AI Principles early and continue to reference them in the narrative to influence the collaborative, innovative culture required to maximise AI potential.

A CEO should follow these four key guidelines to maximise the use of AI by leading the:

• Identification of business value criteria; for prioritisation of the digital backlog of each digital and AI workstream. This prioritises the right use cases and functionality to deliver the required value release or outcome.
• Investment in your data and AI maturity; get the right people, the right data with the right tools and the right controls.
• Setting of the AI principles by which AI should be applied across the organisation. This is an evolving set of rules which determines the outcomes required and provides the permission for the organisation to innovate and embrace AI in a controlled environment.

Whilst you don’t need a new ‘AI strategy’, investing in AI, machine learning, or data science requires careful consideration. As business leaders, by following these four key guidelines, you can leverage AI to drive innovation, improve efficiency, and gain a competitive edge in today's data-driven world.

To talk to us about helping you to get started in AI or to help accelerate your current programmes please contact us.