What Business Leaders Are Getting Wrong About AI

By John Samuel, Founder and Director of Seventh State

Almost three quarters of UK businesses are using at least one form of AI technology, according to the National Bureau of Economic Research – but that doesn’t mean they are getting it right.

The Office for National Statistics reports a 23% rise in use by micro-businesses in the last three years, and even more larger firms have embraced new tools and systems. But there are several common mistakes that many business leaders are making when it comes to implementing AI into their daily operations.

Failing to get the structure in place

Business leaders can easily assume AI can be used as an ‘add on’ to their existing operational structure, neglecting the need for a thorough implementation plan and a well-planned strategy before rolling out new tools.

But without the appropriate structure and systems in place, AI doesn’t work effectively. Machines need clear and accessible data structures to work; otherwise, there is little point in using AI at all.

As much as existing systems need to be optimised before new technologies are implemented, the people within the business also need be ready for the changes. If teams aren’t clear on the benefits, roll-out timeline and correct use of the AI they’re now expected to use, it can become more of a hindrance than a help. Equally, a lack of adoption across departments can see AI failing to produce the expected return on investment.

Deloitte research showed only 16% of businesses had fully redesigned roles, processes and operating models around AI – yet the organisations which had taken a ‘technology first approach’ were 60% more likely to report disappointing results. This supports the theory that without clear organisation and straight-forward implementation, we cannot expect technology to transform business outcomes of its own accord.

Not defining its use

The ONS reports almost 60% of businesses implement AI for the primary purpose of improving business operations. But what does this actually mean? Leaders need to be clear on the tangible benefits, how they will be measured, and who is responsible for monitoring outcomes.

Having this level of clarity around ownership, accountability and expectation from the offset maximises the chance of a successful roll-out, but also puts provision in place for any issues or sub-optimal results to be recognised and resolved at the earliest opportunity.

With such a wide range of AI tools on the market, assessment must be made of which will be the most beneficial before deciding on a strategy. Small amounts of improved productivity through automation may benefit an individual or team, but this must translate to significant gains in value to make the cost of implementing AI worthwhile. It’s reported by BCG Global that 42% of frontline users save at least one working day every week through their technology use – but we see many cases where these time savings don’t convert into business value.

In fact, a Government research paper reports 77% of businesses have not seen a change in revenue since launching an AI tool, while 35% said they had not seen an increase in employee productivity. These figures reflect that solid evidence must be gathered before implementing new technology, and that continued review is needed to identify whether the company is actually benefiting from its use of AI.

Ignoring the value of human judgement

Technological intelligence should not be used as a substitute for human intelligence. A combination of both proves most effective, meaning leaders should not overlook the judgement, skills and knowledge of their teams. While it might be useful for administrative or research tasks, AI cannot capture nuance, adapt in times of crisis or handle context in the way humans can – and, in any case, it requires input and direction to complete its allocated tasks.

Humans and AI aren’t substitutions for each other, so businesses need systems in place that enable both to work at their best. The latter may be excellent at handling data and analytics, but the corporate world is about so much more, and it needs the relationship building, crisis handling, and emotive capabilities of people to function well.

Ignoring the feedback

Too many leaders think implementing AI into their workflows is enough on its own, but tracking performance is vital, and this means continually monitoring, evaluating, and changing course where needed.

Evaluating where these tools are adding value to the business or making the lives of teams easier should be standard practice, with feedback loops established from the beginning to illustrate successes, team responses, and challenges.

There’s no doubt that AI can be hugely beneficial for businesses, but making the assumption that it will be a success from day one without any input or adjustments is a fundamental error.

Getting it right

AI is neither an instant fix not a fully autonomous entity, despite its name. Many of the mistakes that businesses make when it comes to technology centre around trying to remove human input. It is people who must develop a business case and robust strategies prior to implementation, and it is people who must continue to guide and monitor processes post-launch.

Used well, technology can provide immense value. But we must make sure vital steps are not missed, and humans are not engineered out of the situation, otherwise the full benefits will never be realised.