Your InFinance Publication

FINSIA’s InFinance keeps you up-to-date and in-the-know. 

Technology and artificial intelligence to help organisations changing their culture

by Lewis Panther SA FIN | 18 Mar 2019
Electronic data can predict human behaviour and act as an early warning system detecting the kinds of issues raised at the Royal Commission.

The amount of time it takes for people to answer emails indicates just how engaged they are and how likely they are to be making ethical business decisions.

That’s the intriguing claim of Harry Toukalas, who is advising some of the biggest corporations in Australia on how they can fix culture within their organisations.

The former NAB executive and founder of Delphi Bank spoke after the publication of Kenneth Hayne’s final report and urged companies to “think outside the box”.

He said: “I think there's a real opportunity for Australia in terms of corporate Australia, not-for-profit Australia or government Australia to look at what we now know to be the way people think in terms of behavioural economics.”

“I think now is the time for organisations to start thinking outside the box and try to look at ways to approach culture in a different way because whatever they've been doing has been suboptimal.”

For the Blackhall & Pearl co-founder, technology and artificial intelligence are key tools in tackling the problems.

Expanding on the reasoning on why email analysis can be so helpful, he said: “All things equal, the longer you take to respond to an email, the less passion you have for the discussion.”

“The more reminders someone has to give you to respond to their email, the less prioritisation or respect you have for that person. And that's critical.”

“Because if you can imagine, you're the HR department, filling out a culture initiative. And I'm the person that's supposed to be coming on board and I'm taking a long time to respond to that email and you need to send a reminder to respond to that email, it's highly unlikely I'm going to change my behaviour.”

“When you're taking way longer than you should, based on average email response time, that's an indicator that you may not be interested in the issue.

“The other thing we track is the ratio of sending and receiving emails. Where you have a much higher ratio of receiving rather than sending, that means you're not really engaged with the outside world in terms of your organisation.”

“If your inbox is 1,000 email received and 500 emails, or even 200 emails sent, that is a disproportionate ratio of sending versus receiving.”

“We call that balance contribution. Your contribution to the discussion, the more balanced it is, the more engaged you'll be, the more likely you'll be to change.”

While Toukalas is at pains to say that privacy is not compromised, AI gives his team the capability to examine 18 months of emails. That can show an individual usually takes eight hours to respond. But when it comes to anti money laundering, for example, the response time stretches to two days. And that shows a lack engagement, he says.

Even the content of email can be analysed to predict types of behaviour, he adds.

“It’s a very simple thing to do. For an algorithm to find positive, negative words with the right training is pretty straightforward,” he says. 

“Beyond that, it's not just what words you're using but how you're using those words.”

“For example, there's a lot of research now that shows that the absence of a pronoun is an indicator of avoiding responsibility all the way to potentially covering up misconduct.”

“So when we see correspondence with "I" or "I'm" or "I've" missing in a correspondence and people are explaining what happened, that's a very good indicator of shirking responsibility or covering something up. “

“For example, instead of someone saying, "I entered the data incorrectly into the system.", they'll typically say, "The data was entered incorrectly." Or "The system misinterpreted the data." Instead of taking responsibility and saying I.”

“When we first started doing this, there was an aversion to artificial intelligence because of big brother syndrome.”

“Now that was around five years ago. Since then in the last two to three years, we've really seen artificial intelligence or AI become much, much more accepted.”

“People are prepared to at least explore how it could be used and the benefits and the advantages of using it.”

“Provided it is a very transparent process and explained properly and management takes responsibility for it as well. Then there are ways to deal with it.”


Share this