Artificial Intelligence is simply a machine displaying the intelligence level of a human, at least for the specific domain it has been programmed for. One of the earliest precursors was Eliza, a chat program that had responses pre-programmed in. It would pick up words and/or phrases from sentences, and compare the words (or their synonyms) against its response database for an answer. Its responses appeared to be well-thought out, and surprisingly, it offered a human-like level of interaction.
Artificial Intelligence has come a long way since then. Today’s chatbots are self-learning, meaning they improve on their own as more people interact with them. If the response is incorrect, the program makes a note of it and does not make the same mistake twice. It tries another approach the next time. If it is correct, it repeats the response.
Several businesses use now chatbots in customer service. The Royal Bank of Scotland, or RBS for short, uses a chatbot to interact with their customers. It answers a lot of common queries, and because it is a program, the cost savings from not hiring receptionists or contact center workers amount to millions of dollars. The only investment is the cost of developing the chatbot, programming the responses in, and maintenance to see that it is performing as expected.
With standardized chatbot programs now available, development costs have come down. Halfway across the globe, Singapore’s Inland Revenue Authority deploys one such chatbot that opens web pages on its own. ‘Joanne’ then politely lets users know that it is ‘she’ who opened the page, and it contains the information requested. DBS Bank – named the Best Bank in the Asia-Pacific Region – uses a chatbot that makes its users feel like they are talking to a friend. It lets users transfer money, and confirms with them before doing so. So simply typing in “I want to pay $50 towards my phone bill” is enough. There is no need for users to go through an entire gamut of operations anymore to achieve the same, as was the case earlier.
Chatbots are also being used to attract new customers by providing them with a superior level of service. On Facebook Messenger, Activision’s Call of Duty chatbot exchanged around six million messages with users the very first week it was put into operation. It would have otherwise taken a team of 2,500 humans to accomplish what the chatbot managed to do.
Chatbots are but one form of Artificial Intelligence. Machine learning, as it is called, enables them to keep delivering better and better without human intervention. Uber is a case in point. The company has a policy of having only three employees in every city it operates. The self-learning algorithm manages thousands of cars, hundreds of thousands of rides, calculating fares accurately in real-time. About $11.5 billion has been invested in technology since 2010 – and the company’s artificial intelligence program is reaping dividends. The ride-hailing app makes some $10.8 billion every year by way of revenues. A lot of this is kicked back into technology, and the company is now testing self-driving cars and trucks throughout the United States.
Tokyo-based Fukoku Mutual Life Insurance has installed a custom A.I. application that determines policy payouts. It cost them $1.7 million, but the Japanese insurer hopes to now save $1.25 million every year. In other words, their artificial intelligence program pays for itself in less than 18 months. Annual maintenance costs are expected to be in the region of $125,000.
The cost savings come from an increase in productivity. Based on IBM’s Watson, the artificial intelligence program reads medical certificates, surgery/procedure summaries, billing statements, and factors in the length of hospital stays. It then figures out how much needs to be paid to the policyholder, depending on the terms and coverage of his/her policy. Two other insurance companies, Dai-ichi Life and Japan Post Insurance are in the process of implementing A.I. programs for their operations.
Expedia’s artificial intelligence program helps users to find the cheapest rates for hotels and flights worldwide. Three seconds is all it needs to show up to 16,000 options in response to a query. The team at Expedia keeps improving on its algorithm to deliver better results to users. The artificial intelligence program is now capable of generating quadrillions of search results in a single day. AutoTrader uses a machine learning artificial intelligence program to improve their car valuations.
Image processing is also something that Artificial Intelligence programs are capable of. Engie, a French energy company, is using drones to take pictures of their assets, like wind and gas turbines. Their artificial intelligence program scans through the images to find out if there are any damages to the assets. The inspection and analysis of thousands of assets and miles of pipeline is not something that can be easily done by a human workforce. Chief Digital Officer Marc Florette says that timely detection of a crack in a high-pressure gas pipeline can save the company millions of dollars.
These are all examples of how Artificial Intelligence is helping companies grow, adding trillions of dollars in value. At a less subtle level, artificial intelligence is also helping recruiters pick the right candidates to positions, making sense of big data. This is random, unstructured data that is too large to be understood by humans.
It is estimated that Artificial Intelligence will be in widespread usage by the end of the year.