aleutie - Fotolia

Why AI success depends on IT picking the low-hanging fruit

General-purpose artificial intelligence is probably a long way off, but the rapid rate of innovation means off-the-shelf AI can kick-start many projects

New research from consulting, technology and outsourcing services company Capgemini paints a picture of remarkable optimism among heads of business about the benefits of artificial intelligence (AI).

The survey of 1,000 organisations revealed that 78% of them believe AI will increase operational efficiency by 10% and three-quarters expect it to cut operational costs by at least 10%.

The global study is based on a survey of 1,000 senior executives from organisations with revenue of more than £400m that are implementing AI.

Capgemini found that three-quarters of UK organisations implementing AI have created new jobs as a result of the technology.

In the UK, 62% of employees believe machines can greatly augment human output, and 79% think AI will help make complex jobs easier in their company.

Meanwhile, 70% of the survey respondents said AI is bringing new insights and better data analysis to their organisation, and 61% said AI is making their company more creative.

Speaking to Computer Weekly about the results of the study,  Ron Tolido, CTO for Capgemini's insights and data practice,said: “In the short and mid-term, the benefits of AI are based on using off-the-shelf products to apply AI for a specific purpose, then training the AI to make it work.”

Tolido described AI as a “black swan” technology. “We have never seen it before,” he said. “The current state of AI tech is far more advanced than before, such as when I wrote my own Prolog interpreter in the 1980s.”

He said that in the short to mid-term, businesses should not need to focus on mastering underlying AI technology, but rather: “Lots of people need to get used to training AI.”

Read more about AI in business

  • Retailers are beginning to explore how cognitive computing and AI could make e-commerce smarter and more personalised.
  • Pizza chain Domino’s has launched a chatbot that takes orders via its Australian website and mobile app.

So companies can buy something like a chatbot framework and then spend time defining the conversation and dialogue, rather than build a chatbot from scratch. “Anyone can build an Alexa skill,” said Tolido. “You can build on top of very advanced Amazon technology. We don’t necessarily need to care about how it works.”

One company featured in the Capgemini study is Dutch airline KLM, which has adopted an “AI-assisted human agent” model to reinforce its existing customer support staff. “Using voice biometrics, the system can identify over 100 human vocal features to instantaneously authenticate and process a call,” the report said. “The AI agent can also solve customer queries over a variety of digital platforms, adapting the reply based on the inquiry platform.”

Tolido said businesses should also try to understand what a process augmented by AI will look at, such as cognitive AI being applied to process documents. “This requires a lot of creativity,” he said. “It is the augmentation of what a human could not do solely on their own.”

One example cited in the report is investment bank JP Morgan, whose lawyers spend thousands of hours studying financial deals. Now an AI system is performing the challenging job of interpreting commercial loan agreements, taking on a task that had swallowed 360,000 hours of work by lawyers and loan officers, said JP Morgan. 

The AI system reviews documents in seconds and is less prone to error. It has cut down on loan-servicing mistakes, many of which originated from human error in interpreting 12,000 new wholesale contracts a year, according to Capgemini.

When asked in the survey what AI initiatives had been deployed at their organisations, the highest percentage (31%) said customer service, followed by finance (27%).

Tolido described such initiatives as the low-hanging fruit that many businesses can deploy quickly and easily to augment human-centric business processes. However, Capgemini’s report revealed that just 46% of the businesses surveyed are addressing these easy-to-deploy AI pilots.

Among the areas Tolido expects to benefit from AI augmentation are customer service and the EU’s General Data Protection Regulation (GDPR). Deep learning is already being used heavily in areas such as cyber security for intrusion detection, he said.

IT automation

Tolido said IT chiefs should also pilot how AI could be used to improve their own internal IT processes. Such a pilot could be used to demonstrate how AI could be used in the wider business, he said. “At Capgemini, we use AI for service management, cyber security, development and testing,” he added.

“IT departments need to take a good look at themselves, and apply AI at IT processes. These are often people-intensive and sometimes too complex for humans to do themselves.”

For instance, Tolido said there is a strong case to use AI for regression testing in software development quality assurance.

Another rapidly growing area of IT automation is predictive maintenance. Commenting on the latest database technology from Oracle, Tolido said: “Oracle CEO Larry Ellison is claiming it has the first completely autonomous, self-optimising and self-healing database.” He pointed out that many database administrators base their careers on supporting the Oracle database.

The new Oracle product is an illustration of what is happening in IT with AI, said Tolido. “It’s a bit like the ‘turkey at Christmas’ discussion,” he added. .......................................................................................................... ........................................................................

Read more on Artificial intelligence, automation and robotics