Research shows that AI adoption across industry is speeding up – with one report finding that, of over 2,000 organisations surveyed, 47% had implemented it in at least one function, compared to 20% during the previous year.

On top of that, 71% of respondents expected AI investment to increase over the coming year. However, the fact remains that there are still significant challenges to companies wishing to adopt smart, cognitive computing processes into their operations. This is borne out by the fact that in the McKinsey survey, just 21% of respondents claimed they had rolled out AI in more than one process.

My work brings me into contact with organisations at every stage of their digital transformation and AI adoption process. This gives me insights not only into what is working but why barriers or challenges are discouraging adoption. Here’s a rundown of some of the most frequently encountered blockers – as well as ideas being put in place to overcome them.

Cultural barriers

Put simply; this is all about resistance to change. Human beings, it is often noted, tend to be creatures of habit; once we find a method of carrying out a task that seems to get the job done effectively and efficiently, we like to stick with it. It often takes some persuasion before we will see that the disruption and expense that will inevitably be caused by altering procedures or adopting new processes will be worth the overall gains they will bring.

This could be as simple as a disinclination towards what can be seen as “handing over control” – whether that’s directly to machines, or to the human employees who administer the technological infrastructure that makes AI possible.

Often this translates into simply not seeing the need for AI, and an incomplete understanding of the advantages it can offer. And in my experience education is usually the most effective means of overcoming this particular barrier. I spend a lot of my time working with boards and senior leaders to instil an awareness of how the core AI technologies – from natural language processing to computer vision and predictive maintenance – can create efficiencies and reduce costs. Once awareness levels are raised, people are likely to become receptive and engaged with the potential for positive change AI offer.

Fear

Another completely natural and understandable human response. Fear of the unknown, as horror writer HP Lovecraft liked to point out, is the “oldest and strongest emotion of mankind.” And as visionaries including Elon Musk and Stephen Hawkin have pointed out in more recent years, there is a huge amount that is still unknown when it comes to the part AI will have to play in our future.

In an immediate sense, this fear could revolve around a growth in distance between the human workforce and the job they are paid to carry out – decisions made by computer algorithm can be difficult to predict and understand. This leads to a fear that humans are losing control and may possibly no longer be regarded as the “experts” in their field of work.

The middle-term fear is of increasing human redundancy – if machines can carry out the work more efficiently and effectively, where’s the need for humans? Futurologists have predicted that this could go two ways – either leading to a utopian existence where robots fill all our basic needs and humans are free to pursue leisure, intellectual and artistic endeavours, or towards mass unemployment and social unrest.

I believe that instead there is an attractive middle ground (as there is so often), where the solution is to roll out intelligent tech to augment our work and make it better for humans, rather than replace them.

Shortage of talent

This is a real and pressing problem for most businesses wanting to adopt AI as well as move to other data-driven models of digital transformation. Experts predict that when it comes to capitalising on the enormous potential for growth offered by AI, a bottleneck exists due to the shortage of data and technology professionals with the experience and training needed to implement the required infrastructure and organisational change.

Although AI research has been ongoing for decades, it’s only relatively recently that these skills have been in demand by industry. And the massive growth in demand means that those with the abilities are able to ask for very high salaries and prestigious positions within the organisations that employ them. On top of this, those companies that already understand and heavily invest in AI – the Googles, Facebooks, and Baidu’s of the world – are often seen as offering the prime opportunities by those who do have the skillset, meaning other businesses face an uphill battle competing to attract talent.

However, this is likely to be a challenge that will be overcome, eventually, by the old-school economic principles of supply and demand. With data scientist increasingly cited as one of the most exciting (and lucrative) career options in coming years, the pool of talent is sure to grow. Other solutions proving successful include upskilling existing workforces. With the number of AI solutions available “as-a-service” also growing, there will be less need for a workforce that is fully trained in traditional data science, in order to deploy and operate AI solutions to many business problems.

Lack of a strategic approach to AI adoption

In some ways, this is an amalgamation of several other barriers – lack of talent, lack of management buy-in, and a culture insufficiently immersed in the advantages and practicalities of AI and digital transformation. The result is often AI initiatives that aren’t planned at a strategic level, fail to address strategic business objectives and don’t fit within an organisation’s overall plans for growth and business development.

Often the cause here is that, while businesses are broadly aware of the importance of adopting AI technology, and the advantages it can offer, they fail to approach it from a strategic standpoint; this means fully understanding the aims and objectives of all aspects of AI operations, from data gathering to how the insights uncovered are communicated across the workforce and put to work.

The answer to this one is pretty straightforward – organisations must always ensure that a clear strategy is in place before time and money are spent on rolling out expensive and resource-intensive AI initiatives and pilots with no clear understanding of the benefits they can bring. This is another area where I spend a lot of my time, helping companies ensure their AI initiatives are clearly linked to business performance objectives, prioritised by their strategic goals, and where every stakeholder has a clear understanding of what the success – or failure – of an initiative will look like.