The Singapore government's recent announcement of massive investments in AI has again put the spotlight on just how businesses should use the technology, and in particular, how labour-intensive businesses – those that are often the hardest and most costly to automate – can integrate it into their operations.

At an industry panel hosted by Microsoft earlier this week, Certis Group chairman Allen Lew and Seatrium CEO Chris Ong shared the approaches that they are using or planning to use in the implementation of AI.

Certis is a former unit of the Singapore Police Force that subsequently evolved into a commercial security group; Seatrium is the top offshore and marine service provider in Singapore. Both these companies face a challenge common to their industries worldwide: the primary resource they rely on is people, and the cost of labour is going up while the talent pool is shrinking.

So how do companies like these get started with AI?

Look at better uses of scarce resources

In labour-intensive fields, implementing AI needs to be about supplementing the workforce. This may involve putting the technology up front rather than the people – having it do the 'eyes on' or 'hands on' work while humans monitor and intervene only when necessary. Or it may involve using the technology to direct human work more precisely, for example guiding technical experts to the exact locations where they are needed, rather than having them spend extra time to cover an entire work site.

“If you count on having the labour force up front, you will face more and more challenges,” Lew put it.

Give people a simple way to be involved with AI

Training only sticks if people are able to start applying it immediately. Lew shared that at Certis, leaders and managers who go through AI training are assigned two key tasks: firstly, to use AI to solve one simple problem they are facing, and secondly, to find a way of using AI to change the way they are working.

Even board directors do this; he said that during board meetings, the Certis board is now using AI tools to take minutes and create summaries “so that board members don't need to go through 12 pages of minutes”.

Focus on the value of information

Industries that can mechanise and apply robotics, like offshore and marine, have already undergone a “very fierce vertical transformation”, Ong said. But what's still needed is horizontal integration: improving end to end visibility between verticals, so that people are easily able to get the information they need to do their job or collaborate along the supply chain. AI has value in providing that kind of information access and decision support.

Bring along the entire supply chain

Both leaders were firm that AI should not just be about benefiting a single company. The impact of the technology needs to be extended to the entire industry, and one way of doing that is to extend its use along the supply chain.

Companies that are in a position to influence the supply chain, particularly the end purchasers, can do this by focusing on very important areas that affect all the product and service providers, said Ong. He highlighted safety standards as an example of a perpetual subject of interest in the offshore industry, one that can accelerate all the other businesses along the supply chain. First adopters of AI have the opportunity to identify ways of using the technology to boost best practices, and then spread the knowledge to their business partners.

Make sure you have a good feedback loop in place

Finally, any large-scale adoption of AI or in fact any other new technology requires a supportive environment and a culture that focuses on improving people's work experience.

“When you use the word 'transformation' in any organisation, there is immediately a sense of fear,” Lew warned. People will worry for the safety of their job when such things come up, he said, and to keep everyone on board, the organisation needs a psychologically safe environment with a two-way communication process and a very strong feedback loop. Most importantly, he added, leaders and managers need to keep in mind that AI means very different things to people in different roles, and the communication process needs to be adjusted according to who is involved.