While artificial intelligence and machine learning projects increasingly exercise the minds of technology leaders, many organisations find it difficult to scale their initial pilot or proof of concept project widely across the enterprise.

Tim Fountaine, Senior Partner at McKinsey & Company, and head of the company’s specialist advanced analytics arm, QuantumBlack — it is important to recognise that the best approach is not simply to treat AI as a new technology that needs to be introduced, but instead to use the opportunity of transformative technologies like AI to redesign the way the organisation works.

“Perhaps the biggest single thing I noticed is, it’s about mindset and culture. You have to be willing to challenge the way things have been done in the past, be open to new ideas, be willing to experiment, Some things won’t work out so the only way we’ll find new solutions to problems is to be willing to take some risks.”

There are no easy answers, and many factors can stand in the way of moving from pilot to scale. “We can learn a lot from past experiences,” said Fountaine.

By way of example, he described the impact of the shift from steam energy to electricity in the early 20th century.

Tim Fountaine, Senior Partner at McKinsey & Company, and head of the company’s specialist advanced analytics arm, QuantumBlack

“When the electric motor came out at the end of the 1800s, people wanted to replace steam engines with electric motors. The way factories worked was one really big steam engine would run everything in the factory.”

The factory layout was designed around the position of the steam engine, and in the first iteration of change, companies just swapped the steam engine for the electric motor. 

“They got lower energy costs as a result, they got some benefits, but they didn’t really change manufacturing.”

Then, in the early 1900s, new companies worked out they could make smaller electric engines and give one to each worker. 

“They could create new systems and manufacturing production lines to produce Model T cars. They really used the new technology to totally change how things work. And, by the 1930s, many of those original steam-driven manufacturing companies have gone out of business or been replaced.”

Fountaine told Which-50 this is analogous to how companies are approaching AI.

“A lot of organisations are taking these new technologies and sticking them onto the same way that they’ve always done things. It’s an addition to the way they operate, but it doesn’t really fundamentally change [anything].”

This unwillingness to change ways of working is the first thing that gets in the way, he says.

Organisations need to ensure they have the right capabilities, and this is not simply a matter of technical talent, he said. “Your frontline salespeople — your marketers for instance — they need to understand that (change) too.”

That naturally leads to different ways of working, he said. “You can’t do it with data and analytics people sitting off in a room by themselves. They have to be working with people in the organisation solving problems in that cross-functional way. That can be really difficult, in a traditionally organised function-based organisation.”

Previous post

Attending to Environmental, Social and Corporate Governance Must Include Digital Infrastructure

Next post

Pandemic drives Canva’s $US15Billion Valuation