Think Paris 2018: when AI reinvents business and management
A smart chatbot now handles half of all customer conversations at Orange Bank. Crédit Agricole CIB has automated the job of reading and analyzing corporate annual reports. Mana, an environmental association, effortlessly pools 90% of all the available information on the environment, allowing it to concentrate on the task of analyzing climate change. Each of these cases was presented at Think Paris, the annual IBM conference organized on 9 October at the Carrousel du Louvre in downtown Paris. The 2,000 attendees heard about the practical new applications developed under the Watson artificial intelligence (AI) program, IBM’s future AI investments in France and the AI projects being taken forward by emlyon business school, presented by none other than Bernard Belletante, the school’s MD. Read on to catch up on an event that showed how AI’s business impact is gathering speed. Better be quick, though, as this article will probably be out of date in six months!
AI, a smart business accelerator
A very human chatbot
Setting the science fiction aside, AI is discreetly shouldering more and more low value-added tasks. This point was illustrated at Think Paris’s opening presentation, which introduced the audience to Djingo, the virtual assistant used by Orange Bank, an online bank. In actual fact, Djingo, which is capable of text and voice interaction, is a white label version of Watson, the AI program developed over the last decade by IBM — that’s right, the one that passed the Turing test in 2011 by beating a human in Jeopardy, an American TV quiz show. But Watson/Djingo has come a long way since then. It now handles 50% of exchanges between customers and Orange Bank, compared with 36% when it was first launched. Able to answer 500 questions in natural language, Djingo only gives way to a “real” advisor when the conversation becomes too complicated. André Coisne, the former CEO of Orange Bank who has been tracking the project’s introduction, calls it the ideal solution to optimize customer interactions.
AR-based tech support
While AI is helping to simplify and automate customer relations, it also has practical applications in providing support to technical staff. IBM demonstrators showed how a blend of visual recognition, machine learning and FAQ processes can help to create a virtual repair solution. A technician working on a printed circuit board or motor merely needs to hold up his or her smartphone: the program will immediately identify the faulty part and provide an augmented reality (AR) explanation of how to replace it, showing which wire to disconnect, what steps to take, etc. The possibilities for this kind of application are endless.
Participants also heard about how Crédit Agricole Corporate and Investment Bank is using machine learning and data analysis to conduct close readings of financial reports. Watson can analyze and summarize documents of over 100 pages in barely three minutes, where a financial analyst would take four or five hours. The AI solution is able to rank information, highlighting countries where a firm does business, identifying partnerships and links to suspicious companies. “Watson’s machine learning capabilities are being harnessed to enhance risk management, compliance and fraud prevention”, explains Jean-Philippe Desbiolles, Vice-President Cognitive Solutions IBM Watson. Here again, the aim is not to replace human beings, but to let people concentrate on high value-added tasks, including actual financial risk analysis.
AI is now so deeply anchored in daily life that even NGOs working to promote sustainability have started using it. Mana Community is an association that identifies environmental risks associated with business practices. Collating all the information available on search engines and social media and making it available to large groups is a mammoth task. “Pre-analysis of the data by Watson has allowed us to cut data processing by 90%”, says Mana founder Kiti Mignotte. That leaves team members free to concentrate on the remaining 10%, which adds up to appreciable time and efficiency savings.
AI: better than an antivirus when it comes to cybersecurity
It may be a less intuitive application, but cognitive learning algorithms could be useful allies when it comes to cybersecurity. In an area where experts are in short supply, Watson Security is assisting the people in charge of network security. The software draws on a gigantic database of all current threats to quickly spot attacks, identify viruses and provide timely counter-attack hypotheses to keep networks safe. Once again, the cognitive learning solution merely complements the human element; IBM experts and researchers are constantly watching to identify new vulnerabilities and prepare antidotes for Watson to use. When you consider that one in eight users worldwide ends up clicking on a phishing link, the extra security should come in very handy.
How AI is changing the ways we learn
AI is revolutionizing the ways that companies operate and transforming corporate cultures. So you’d better make sure you are getting tomorrow’s managers ready for these changes. This was the idea behind the Think Paris presentation by Bernard Belletante, Managing Director of emlyon business school. We have to rethink everything, learn in new ways and become mobile. “We’ve moved from a stock-based approach, with profs, programs and classrooms, to a flow-based approach”, says Bernard. AI is changing our vision of schools and companies, and outlining the jobs of the future. Which is why IBM and emlyon business school have launched a successful tech partnership that will transform the student journey. The school has created a system that can predict job trends over the next five to seven years and then introduce the educational elements that students need to follow to be job market-ready. Bernard describes it as a “vast GPS for skills that is capable of integrating multiple unknown factors, including jobs that don’t even exist yet”. In short, the system can anticipate how the world will be reshaped by AI — using AI! This is a huge challenge for the future.
Enhanced student profiling
In the meantime, a first AI-based test was successfully completed on the emlyon business school campus during the 2018 back-to-school period. Over the four intake days, 1,450 students were asked to enter their interests and profiles in a specially designed program. An AI algorithm was then used to separate them into 220 working groups, which were tasked with suggesting ideas for business start-ups. Of the brand-new start-ups, 50 were selected to go on to the incubator stage. Bernard explains: “Student profiling made this test possible. Processing the data would have been impossible without AI.”This experiment is just the beginning.
What can we expect from AI in the next five years?
One of the highlights of Think Paris was the presentation by Haig Peter of the IBM Research THINKLab about major AI and technology innovations that are going to have a direct impact on society and the business world.
Quantum computers for AI?
While research into quantum computing has been underway for 40 years, the first quantum computers are expected to come out within the next three years. IBM has already provided its researchers with a 5-qubit cloud-based quantum computing system. “It won’t solve all the issues, but the enormous increase in computing power will make it possible to handle complex and intelligent management operations — for an airport for example”, says Haig. IBM’s quantum computer has already been used in nuclear physics. Expect quantum computing to be one of the big IT trends in 2022.
Harnessing AI to protect the oceans
Haig also spoke about how AI could be used to monitor and protect the world’s oceans by deploying microscopic AI-operated robots in plankton to monitor water conditions, identify areas of plastic build-up and track plankton developments.
Keeping AI neutral
Last but not least, the rise of AI could have undesirable effects through an explosion in AI bias. To give an example, for three years, Amazon employed an HR pre-screening algorithm whose decisions turned out to be biased against women. By relying on data on the existing workforce, which was mostly made up of male workers, it systematically rejected applications from women. “Developing systems that learn without prejudice is crucial if AI is to take truly fair and effective decisions”, says Haig. Making sure that AI is neutral and impartial is a huge challenge.