The various impacts of Covid-19 have the potential to spur uptake of artificial intelligence, but companies need to get the right culture in place to best take advantage of it, says a senior Microsoft executive.

Mitra Azizirad, a 28-year veteran of Microsoft and now Corporate Vice President, AI and Innovation at the Redmond, Washington-based software giant says AI adoption was building pace, driven by three key areas organisations were prioritising.

They included business process optimization, improving employee productivity and boosting customer service. The latter was a particular priority as the pandemic forced changes in how businesses dealt with their customers. 

“Now more than ever, airlines, public service agencies and many other companies are facing an overwhelming amount of calls from customers,” said Azizirad who last week delivered her first Linkedin virtual keynote from her home in Washington, complete with an AI-powered background driven by a Microsoft’s Kinect motion-sensing camera.

Health bots take off

“We see more organizations turning to AI to be able to immediately respond and deliver quick impact for the customers, their patients and citizens in response to the COVID-19 pandemic,” she added.

Around 1200 Covid-19 self-assessment bots had been built on Microsoft’s healthcare bot platform, reaching over 18 million people across the world. 

When it came to employee productivity Azizirad was seeing a renewed push by businesses to move from “manual rule-based systems to human supervised automated engines”.

She pointed to global news wire service Reuters, which recently implemented an AI recommendation engine to match relevant videos to articles it journalists publish. 

“With this approach, Reuters has been able to increase total video views by 4 per cent while also driving the average completion rate above 80 per cent,” said Azizirad.

On the business processes front, Azizirad pointed to Microsoft customer State Farm, a major US vehicle insurer. 

“A powerful example is how they’ve used AI and machine learning, greatly reducing the time to settle an auto damage claim from 15 days to as little as 30 minutes, which is a win-win for both the business and for their customers,” she said.

What’s the strategy?

Organisations that truly took advantage of AI across the board made their decisions in new ways, explained Azizirad. They drew insights from large sets of data, giving them scope to achieve more productive results.

But they also had to take the time to get “super clear” on their strategic imperatives across the entire organisation.

“It is super expensive, we all know, to experiment with projects and multiple proofs of concept with limited application or ROI,” she said.

“So be super selective about where teams should be spending their time. The old adage to slow down in order to speed up has never been more true,” she advised.

Microsoft had developed its free AI Business School platform, which offers masterclasses in all aspects of AI planning and implementation, to help organisations adopt the technology. One million people have now completed courses on the platform, development of which has been informed by feedback from Microsoft’s own customers, including the cases where AI projects hadn’t gone to plan. 

“My team and I have spent the past couple of years crisscrossing the world working closely with business leaders, AI experts and professionals, and having deeply candid conversations about their experiences, about what worked and quite frankly, what didn’t as they began using and rolling out AI,” said Azizirad.

Ultimately, having the right culture in place was crucial to the successful application of AI in business. 

Four ways to instil an AI-drive culture

She said there are four aspects to having the right culture for AI

Be data-driven:

An organisation needed to be able to extract and analyse data over its “entire data estate, no matter where the data resides”. Increasingly, that also meant accessing data in real-time to make decisions at speed. Having silos of data that were crucial to decision making but couldn’t be accessed quickly was a barrier to effective use of AI on an organisation-wide scale.

Include employees:

It seems simple, but getting buy-in from employees wasn’t always front of mind when it came to introducing AI technology. 

“It’s about enabling all employees to participate, contribute ideas, ask questions, so every layer of the organisation is represented,” said Azizirad.

An inclusive approach would lead to more diverse ideas, better solutions and more motivated employees, she argued.

Implement AI responsibly:

Every organisation had an obligation to address the ethical challenges use of AI in the organisation presented. This could be to do with issues of data privacy, bias in algorithms or lack of transparency over how AI systems worked. A governance model that operationalised responsible AI was crucial. 

“You need tools that detect bias and data model transparency and explainability, which is critical in making decisions fully transparent,” said Azizirad.

Leadership: 

Finally, for AI to have a transformative effect, it needed buy-in at every level of the organisation, including support at the “uppermost levels”. 

“It needs to be ingrained into the fabric of how decisions are made, and how employees are managed and rewarded,” she said. 

“It is this marriage of cultural transformation with the adoption of technology that has a true force multiplier effect.”

Microsoft’s Responsible AI Resource Centre was a section within the AI Business School offering advice on how to approach the ethical and governance issues associated with responsible use of the technology.

On an operational level, accelerating uptake of AI at scale required a strong machine learning operations practice. That meant meaningful collaboration between data scientists, machine learning engineers, software developers and other IT teams to manage the AI lifecycle, connect it to the business and make sure it was done responsibly.

“In my experience, this is an area where I have found most customers expend the least amount of time and it’s one of the most important areas,” said Azizirad.

 The pressure was enormous to answer tough questions as Covid-19 wrought destruction on the global economy but with the right culture in place, AI could play a valuable role in decision making and helping organisations big and small achieve their goals.

You can watch Mitra Azizirad’s full LinkedIn virtual keynote here.