If only AI was the panacea it aspires to be.

Every business process, communication, and interaction automated: zero human intervention required; performance beautifully optimized; errors and misinterpretations seamlessly eradicated.

If that is where we are all ultimately headed, we are currently – and many may say thankfully – not there yet.

Recent, well-publicised controversy surrounding dubious, AI-generated content on Microsoft’s MSN.com news website shows just how important it is for humans to maintain a controlling hand.

Inaccurate news stories published on the site ranging from Covid-19 conspiracies to US President Joe Biden falling asleep in public have led to concerns being expressed by major global news outlets that Microsoft’s replacement of skilled journalists with AI technology is a dangerous step too far.

Indeed, this month, in response to the criticism, Microsoft issued a public statement saying it is committed to addressing the recent issue of “low quality articles” on the site and that it is working with content partners to ensure its “standards are met”.

But what does all this mean for enterprise?

Well, it supports the case for Augmented AI – the blending of human interaction with powerful machine learning engines in ways which not only validate and classify communicative data, but which also provide those engines with crucial verification updates. In short, keeping it all real.

“AI is obviously positively transforming the way in which the world works, but it’s not a panacea – for some types of businesses, giving up full control to AI is not wise or accepted,” says Chris Stapenhurst, Senior Principal Product Manager at leading data management and protection experts Veritas, whose Augmented AI-powered ‘intelligent review’ feature is baked into its smart financial services communication surveillance solution.

“Banks and financial institutions must ensure they are doing business in a compliant way. Deploying AI to capture and analyse communications across all channels and then to alert the organisation of issues which present risk is vital. However, combining that functionality with human reviewers who continuously train the AI as part of their review process by highlighting what’s really important is even better as it enables checks and balances along the way.

“Prior to deploying Augmented AI, one of our large banking sector customers had a team of 40 people reviewing communication between brokers, traders, financial advisors and customers. Turning on a machine learning engine with Augmented AI enabled them to reduce that team to just four.

In addition, they used to review something like 80,000 communications every week but now review about 12,000, and they have tripled the number of detected violations. That’s 36 less people and 300% more efficiency.”

The Veritas surveillance solution captures all omnichannel communications and stores the data in line with financial sector regulations. It then classifies the interactions and enables quick and easy searching and sampling by human reviewers looking for signs of non-compliant language or actions.

A reviewer-selected ‘content snippet’ indicates to the machine learning engine that a particular portion of the communication contains text of interest, for example zeroing in on a particular sentence within an email. The solution’s ‘continuous active learning’ feature tasks the machine learning engine to watch the reviewers as they mark items responsive or non-responsive. Consequently, when new items come into the system, it is able to better-predict whether reviewers should look at them.

“Financial services regulation is evolving all the time and so violations are changing all the time,” says Stapenhurst.

“Also, the channels via which we communicate are changing all the time, meaning the world is becoming more and more complex. Trying to protect against risk by doing simple keyword searches and samples just doesn’t fly anymore. Deploying Augmented AI means the machine learning is very sophisticated, very current, and stays up to speed.”

For Veritas and its reseller partners alike, the pre-sale conversation is a compelling one; the return on investment easy to articulate.

“Yes, banks are interested in saving costs, but they are much more interested in reducing risk of non-compliance and the resulting penalties,” says Stapenhurst.

“The world’s large financial institutions expect to have to pay an estimated combined total of $100m in fines every 10 years or so. That’s a very compelling reason for them to do everything possible to remain as compliant as possible.”

To learn more about how Veritas can help your and your customers’ businesses leverage the power of Augmented AI communication surveillance, click here.



from UC Today https://ift.tt/zmQBkZF