Don’t implement AI technologies just to keep up with the Googles of the world
6 min read
17 November 2017
Considerations for any growing business looking to improving the way things are done through AI technologies and automation.
Everyone is talking about the “AI revolution”. From machine learning and cognitive computing to smart robots and drones, AI technologies will be “at the centre of most disruptions over the next ten years”, according to Gartner.
Businesses with AI technologies will be able to “harness data in order to adapt to new situations and solve problems that no one has encountered previously,” says Gartner. And those that successfully apply it, Accenture adds, could boost profitability by 38 per cent by 2035.
Companies are thus rushing to get involved – often without any over-arching plan – and large enterprises such as Google are investing billions in AI-related R&D or acquisitions. Some industries are further into adopting AI successfully: telcos such as Vodafone, for example, use “intelligent assistants” to improve service levels.
But the majority are still unsure where to start. In a recent poll to find out how bosses were tackling AI technologies and automation, we found that 50 per cent were primarily interested in how it could add value or improve service. We also found the majority lack the in-house expertise to implement AI. That is a common concern whatever the business’s size.
This is a revolution precipitated by technology, so it’s inherently iterative and hard to ignore. But you don’t just want to implement AI to keep up with the Googles of the world. It makes sense to approach AI projects strategically, with a structured programme.
Start with now
It is vital to have a detailed picture of how the business runs today, defining current processes and running costs, the value to the business of repeatable processes, the impact of potential failure and so on. This will help identify where the business might get the best return from any AI investment.
Ask what an AI project might achieve and what the desired outcomes are – dig deeper than just “cutting costs”. Is it to reduce expensive errors? Is it about automating repeatable processes? If so, what level of judgement do these require? What about the existing IT infrastructure – is it able to support AI technologies?
Build a centre of excellence
Most businesses hand off responsibility for AI to tech teams. This is a mistake. AI implementations should be operationally-led, with a senior sponsor setting specific goals and deadlines. So put together a multi-disciplinary team to implement the AI strategy. This should obviously include both IT and business process architecture expertise, but it shouldn’t stop there.
The more diverse the team is and the more of the business it draws from, the more it’s members will be able to learn from each other. Some input from tech providers is fine, as long as they are not the only voices in the CEO’s ear.
Take incremental steps
The outcomes are often unknown when it comes to new developments in AI, so try beta testing for a few hours at a time. Measure, tweak and fine tune these experiments quickly based on feedback. Customers won’t mind, especially if a bot solves their problem as efficiently as a human.
Capitalise on size
Costs are always a concern, but small businesses need not be at a disadvantage. Starting with a blank page, entrepreneurs and SME bosses may actually have the edge: there’s no need to try and unpick legacy systems, and they can move with greater agility during pilots.
Arguably, it is growing businesses that will identify the most real gains from AI, too. Without big budgets for pet projects, they have no choice but to focus their investment on something tangible and operationally justifiable.
See the big picture
Look at AI holistically, rather than as a point solution. Consider how the introduction of new technology might impact the shape of the workforce and the structure of the organisation, both in the short and longer term. How will people respond? It’s unlikely that swathes of people are going to lose their jobs – human roles are more likely to be enhanced, not replaced, by automated systems.
Nevertheless, employees may be fearful. It’s important to reassure and to inform: articulate the company’s strategy for introducing automation or AI, be honest about the short-term and longer-term implications for employees. In some industries, the organisational structure will change more radically than others, but even the earliest adopters won’t transform overnight.
It might be worth considering getting independent analysts in for a briefing. Less formally, use networks and corporate outreach programmes such as Grant Thornton’s Business Lounges to exchange experiences and make contacts. Go to forums and use every opportunity to learn more about what’s going on in AI.
Don’t just talk to technology vendors – connect with other organisations that are starting to test out AI technologies. Learn from them. And identify someone to act as a ‘horizon watcher’ and to be the repository for new ideas within the business.
Stuart Jansze is a partner in the technology, transformation and change practice department at leadership resource consultancy WBMS, part of The Wilton & Bain Group.