Augmentation and Innovation...Not Just Automation

AI is regularly plagued by the headline: ‘robots are going to steal our jobs’. The fear felt by workers in roles which are looking to be automated with intelligent machines is of course understandable and justified, but there is much more to the AI conversation if instead of replacement, we consider what can be added to jobs done by people.

Cyrille Bataller, Managing Director – Applied Intelligence at Accenture, spoke to us about how AI can help us transform the way we operate, by improving outcomes for organisation and individuals alike.

There’s much that AI and robotics can already help us with. Bataller explains: “You can have robots read and understand content, such as scanned documents or images or videos, phone calls etc. Also dialogue in natural language, speech or text, to interact more naturally with people so robots can integrate themselves into the workforce.”

Automation though has, in the past, only been able to go so far. ““Automation is what everyone immediately thinks about when you mention software bots, and there is indeed a whole range of repetitive tasks that could not be automated before,” Bataller explains. “Even robotic process automation is very basic in what it can do; as soon as there is natural language text or images that require categorisation or that are used to make judgement calls, traditional RPA bots can get stuck, and therefore, even though they are high volume repetitive activities, they could not be automated before.”

Adding in the human element is key, then, in really realising the true potential that AI and the automation that can come with it, might have. “When you augment with AI, you can suddenly deal with much more complex repetitive tasks, so you can automate a lot more and generally what you automate is what is less interesting for people, so you are doing them a service as you are doing the tasks better faster and cheaper, freeing up people to focus on the more important side of their job.”

The focus on, and fear surrounding, automation capabilities can be understood – the potential in turning monotonous manual work into automatic digital processes is indeed huge and there are a lot of human tasks in their current state that may be made redundant – but having this as the sole focus is not only fear-inducing, but missing some of the larger possibilities that AI and robotics bring when combined with human capability.

Bataller sees this coming to life in three ways, as a ‘virtual workforce’ to enhance businesses as well as worker experience and expertise. First there’s the ‘basic’ automation of high-volume, complex, repetitive tasks which most people think of, but the second and third are not only less discussed but far more compelling. “The second part is augmented decision making… What I call collective experience: the 10 top experts in the company and their 100 best individual decisions are used to train an algorithm that will create a prediction. You then give this to the junior employees to help guide them towards what is likely the best decision to make. And by providing data insights: humans are not good at crunching high volumes of data but machines are. Relying on a machine to show trends in a large data set helps us make the right decision faster.”

“The third aspect is scaling new innovative business services because we don’t have the capacity, for example, to watch every CCTV camera in a city, but when we have bots that are cheap, cloud sourced and efficient, and can mimic what a human would do watching a camera – e.g. counting people and vehicles, then you can suddenly innovate as you can have 30,000 bots watch 30,000 cameras 24/7 which can alert you if any problems happen. Then there’s a need for more employees, not fewer, to handle the alerts – thus delivering a better smart city service to citizens.”

This deeper consideration of AI and robotics, as an augmenter and innovator as opposed to just a replacement, can open up inspiration and broader thinking around solving business problems. Cyrille gives an example: “There’s an insurance company that receives 4 million customer emails per year in 3 different languages. They are keen to handle these emails appropriately and so have asked us for a target of 98% accuracy in our categorisation. There are 400 categories in the types of emails received, and they had a team of 80 full time people just routing these emails. So we’ve built a sophisticated categorisation model with multiple layers in order to route these emails with high accuracy. To achieve 98% accuracy, we only automate about 25% of the routing, the rest are a bit below this confidence level, and is double checked by staff. If they relaxed the accuracy target to 95% we could maybe handle around 70% of the emails. This means we are freeing up around 30 out of the 80 full time employees, and these employees can now respond to emails instead of merely routing, to provide a better service to customers.”

Bataller believes we should embrace the power of AI and robotics for this kind of business operations, and move beyond the conversations of fear. “AI has a transformative power: when you suddenly have access to a human-like, cheap workforce that you can provision very fast, you have transformation potential for your organisation and disruptive potential for your industry,” he explains. “You should disrupt before you can be disrupted.”

Meet Accenture and the entire AI community at this year's World Summit AI in Amsterdam. View the full programme, speaker line up and book tickets by visiting worldummit.ai

World Summit AI 2019
October 9th-10th
Amsterdam, Netherlands
worldsummit.ai