As we have discussed already, the need for organisations across sectors to move beyond executing small-scale AI projects and become truly AI-enabled is pressing – and one that will play an increasingly vital role in their ability to gain a competitive advantage in the future. In this chapter (as well as in those that follow), we therefore focus on how they might go about doing this.
Firstly, though, we need to be absolutely clear on why. After all, adopting AI at scale is a significant undertaking; one that will affect people and business units at all levels and across multiple areas of the organisation.
It is understandable, therefore, for leaders to ask: what is the benefit of striving to expand the use of AI even further?
The answer is compelling: organisations already making progress on their journey to implementing AI at scale are performing 11.5% better than those who are not, a figure that has more than doubled since last year. In particular, they are more productive (11%), show higher profitability (12%) and experience better business outcomes (11%).
This offers a huge boost forward for any organisation, carrying not just the obvious impact on its bottom line but also significant advantages when it comes to stealing a march on rivals. As Microsoft UK Chief Executive Officer, Cindy Rose, explains: “There is now a clear link between an organisation’s full deployment of AI technologies and its ability to gain and retain a competitive edge.”
Organisations already scaling AI are performing 11.5% better than those who are not.
Similarly, AI-advanced organisations (those that are successfully employing the technology at an organisational level rather than just a local or departmental one) are more agile than those that are experimenting with it, meaning they are better equipped to respond to customer and employee needs, changes in technologies, or market conditions. Meanwhile, the ones experimenting are more agile than those not investing in AI at all.
Strides not steps
While more than half (56%) of UK organisations are now using AI to some degree and the number with an AI strategy in place has more than doubled (from 11% in 2018 to 24% today) is encouraging, it does not mean progress is necessarily happening quickly or expansively enough.
Put another way: first steps should not be mistaken for giant strides. Of all the business leaders we surveyed, only 8% classified their organisation as Advanced AI users while nearly half (48%) currently remain in the experimentation phase. (See Figure 4.)
Figure 4. How UK organisations are progressing on their AI journey
Advanced-AI organisations are classed as those in which AI is involved in most things they do, while those labelled as experimenting are using it only in discrete business areas or functions. Meanwhile, given the pervasiveness of AI in applications and software services, those who rate themselves as either doing nothing or don’t know are most likely not doing any conscious implementations rather than failing to use it at all. Opposite are how UK business leaders are split across those categories.
Moving from the 48% to the 8% is a holistic process. Indeed, the best way for organisations to scale AI is the same way they should any other aspect of ongoing digital transformation. Namely that it is iterative and far more than a purely technical project owned and overseen by the IT department. It is something in which the entire organisation must feel consulted and engaged. As Microsoft UK Chief Operating Officer, Clare Barclay, points out: “The more organisations embrace the need for holistic cultural transformation, the faster they will be able to scale their use of AI. Leaders must take action now to embed diversity and inclusion as well as ethical principles into their AI strategy, bringing to bear the skills of all and ensuring no one gets left behind.”
Start with the problem
Equally key is knowing where to start. That is, no AI project in the experimentation or implementation phase, is likely to get very far unless everyone involved is crystal clear on the actual business problem being solved.
This includes having a clear view on what the AI is expected do to, along with what level of resources are required to introduce, manage and measure it.
Indeed, as we see in the Box Out on the following pages, being clear on both purpose and function will help prevent an organisation from either under- or overextending itself while also mitigating the risk of it falling into the adoption chasm between experimentation and implementation.
Taking the time to identify the issue, set the right course of action and improve progressively to achieve their goal is the smartest way forward.
Technically correct
Similarly, something all advanced AI organisations tend to have is a strong data strategy – accompanied by the tools and capabilities needed to deliver it.
The importance of this is recognised by business leaders too, with 43% agreeing that preparing usable data represents their biggest challenge to scaling AI. Remember also that Analytics and Big Data technologies currently top the list of AI applications being used by businesses in the UK (see Figure 2, page 8).
This need for companies to get their data house in order is true across all sectors, with experts from the fields of finance, healthcare, retail and manufacturing united in seeing it as a critical component of any AI scaling plan.
It is what Dr. Lee Howells, Head of AI at PA Consulting Group, which specialises in management consulting, technology and innovation, refers to as the “critical mass” – the big, evolving picture that can uncover game-changing relationships between information sourced from different parts of the organisation, then use them to power new, company-wide AI applications.
The leadership challenge
Of course, moving from experimenting with AI to implementing it at scale is a complex process. One that cannot happen overnight and that brings with it a number of significant, albeit surmountable, barriers to success.
Chief among them is the pressure it places on leadership. Yes, in a strong organisation, all staff have a responsibility to be curious and proactive about driving positive change, especially, as we have seen in previous years, when it comes to digital transformation.
Yet as the people tasked with setting an organisation’s strategic direction, leaders are being required to absorb a great deal of new information about the capabilities of AI and then quickly ascertain exactly what role it can and should play within their organisations.
Positively, far more leaders want their firm to be seen as trailblazers in AI innovation, than this time last year. They now need to back that up with action, in particular by investing more in their own AI education and training so they can then pass on that same knowledge and progressive mindset to employees.
Yet our research also reveals that many UK leaders lack an understanding of how AI can be used in a fair, responsible and effective way. Currently, just over a fifth (21%) say they have completed training in how they can use AI in their jobs while around two-thirds (63%) do not know how AI works out its conclusions. Overcoming this will be a crucial factor in whether they are able to implement AI quickly, effectively and responsibly.
OceanMind is a not-for-profit organisation that empowers enforcement and compliance to help governments and the seafood industry to protect the world’s fisheries. Here, Chief Executive Officer, Nick Wise, explains how the company is using AI to solve a very human problem.
“A big part of what we do is looking at the movement of fishing vessels around the world. Of those we are able to track (eventually 3 million), each one produces somewhere between tens and thousands of data points per day. In the past, all this data had to be reviewed by a human who would look at a particular vessel, understand its movements, put those movements in context with regulations and decide whether or not there was something suspicious to follow up on. For countries like Thailand, with thousands of vessels, that is just an insurmountable problem. You simply cannot get enough people with the right expertise to do it. So, we trained a machine to understand some of the key knowledge among our employees. This meant it could perform an initial sift of the data before it was passed on to us humans. So, whereas in the past a team of people would have to look at the data and then make trade-offs and choices about what to analyse based on the time available, now, with AI, it is possible to sort the data much faster according to its characteristics. Based on what the AI filters, the teams can then spend their time looking only at information flagged as anomalous or important.
Using the technology has already saved countless hours and helped us be far more accurate and effective in selecting the vessels and data we investigate. And the reason it has worked so well is we started with a specific problem for the AI to solve rather than the other way around. For any organisation, do that and what can be achieved with AI is virtually limitless.”