JRB - Fotolia
Chetan Dube left New York University, where he was a professor of IT, in 1998 convinced that there could be a more efficient way of cutting costs than replacing human labour with cheaper human labour. He began a mission to find a way of replacing human labour by cloning human intelligence.
Some 17 years on, he believes the world is approaching an industrial revolution driven by artificial intelligence (AI) that will see an unprecedented rate of change.
He says many people gave up on developing AI capable of carrying out the work of humans and he viewed the drop in the amount of research being done as an opportunity to fill the gap.
“AI went through the classic Gartner hype cycle of a peak of heightened expectations followed by a trough of disillusionment,” says Dube. He says early AI researchers were confident and, in the 1960s, Herbert Simon predicted that machines would be capable of doing any work that man could do within two decades.
But one of the founders of AI research later said the problem was much harder than anticipated, says Dube. “This trough of disillusionment led to a period of 'AI winter', when there wasn’t much research or funding for AI-related domains,” he adds.
After leaving university, Dube formed IPsoft, an enterprise IT supplier that has become the vehicle for his mission.
At IPsoft, Dube has built a large business selling autonomic computing systems, but perhaps his greatest achievement is the creation of Amelia, a robot that doesn’t just automate business processes or IT, but is capable of reasoning and which, says Dube, will become “the best friend and the most faithful servant that man would have known”.
Amelia, as Dube calls the software platform, has an understanding of the semantics of language and can learn to solve business process queries like a human. It initially learns using the same manuals as humans – it can read 300 pages in 30 seconds – and then learns through experience and by observing the interactions between human agents and customers.
If Amelia can’t answer a question, it passes the query on to a human, but remains in the conversation to learn how to solve similar issues in future.
It understands 20 languages as well as context, and can apply logic and infer implications.
The software is already used for services such as technology helpdesks, contact centres, procurement processing and to advise field engineers, among other business processes.
Smart does not always mean smart
But Dube says it is not easy to sell AI to senior executives, who are becoming confused by the many forms of IT that are described as smart – wrongly, according to Dube. He says there is too much “marketing fluff” about smart systems today, which is underselling “truly smart” machines, such as Amelia, in business.
“Many devices today are referred to as smart, but unless they are capable of learning and making decisions based on that, they are not in fact smart,” says Dube. “The current fluff around everything that is smart is just a marketing distraction rather than anything to do with intelligence.”
This is undermining confidence in the possibilities of AI, he says. “The question that the market is increasingly asking is: ‘can a thing be classed a smart if it cannot match the intelligence of a 10-year-old?’.”
This is causing confusion in the enterprise sector, with businesses disappointed by the promise of smart systems that are not actually smart, says Dube. “A customer told me his company had a smart agent in place for 12 months but it still couldn’t do the job,” he adds.
Dube says true cognitive systems like Amelia are rare, but are being used and developed in big businesses – and Amelia has many suitors in the enterprise space.
Read more about artificial intelligence
- An open letter signed by more than 12,000 technology experts calls for a ban on artificial intelligence (AI) to manage weapons “beyond meaningful human control”.
- Artificial intelligence in the enterprise isn't some far-off science-fiction film fantasy. It's already here, and it's time for CIOs to judge its business applications.
- Socially aware general-purpose artificial intelligence in the form of a dog could be the ideal form factor to take over the world.
- Greater automation means the boundaries are moving and more jobs could be taken over by a computer.
But Dube is protective. Enterprises that want to use the system cannot just come up with the money, but have to meet other criteria. Currently, 162 Global 2000 companies have said they want to use Amelia, but only seven currently do because Dube is very protective to avoid any disappointments.
“We are only integrating with companies that want end-to-end digital transformation,” he says. He doesn’t want Amelia to become just another smart machine, but wants to ensure the system learns and becomes better at what it does and is part of larger projects.
“Amelia is capable of transformational impact,” he says. “Allowing her to transform a service desk enables the enterprise to realise savings on average above 60%, while improving customer experience by bringing down the time to resolution of incidents. Extending her reach, in a phased manner, to allow companies to move towards a full digital transformation for competitive gains is a natural derivative of her prowess.”
One customer using Amelia is a large US media company that has already realised the huge advantage of the system in a call centre agent context. The company, which Dube would not name, receives about 65,000 calls to its contact centre every month. It was taking an average of 52 seconds to answer a call and an average of 18 minutes for a satisfactory resolution.
When the company moved to Amelia, the calls were connected automatically and it was taking an average of four and a half minutes to reach a satisfactory resolution.
Finance, media and healthcare are the main industries that began using Amelia. “She is getting very competent in these verticals,” says Dube.
The next wave
Telcos will lead the next wave of adoption as they attempt to add value in the face of competition from over the top (OTT) suppliers. “OTTs are killing telcos,” says Dube. “They are using the assets of telcos themselves to, ironically, cannibalise the telco’s revenues.
“Telcos need to better monetise their assets by providing value-added services, such as providing financial and retail advice, rather than battle the OTTs on commodity data and voice services.”
Amelia learns as it goes and this benefits all customers because if it learns tasks where there is no differentiation between customers' needs, it will pass the benefits on to others. But to protect competitive advantage, proprietary information that gives a customer an advantage is not used by Amelia beyond the customer that owns the information.
IPsoft has helped many customers reduce the amount of work they offshore to low-cost locations such as India, which have been used to cut the cost of operations such as BPO. AI has enabled businesses to replace offshore staff with robots.
But it doesn’t end there. Artificial intelligence platforms like Amelia will take over a wider range of roles as they become smarter.
Read more about IPsoft
- IPsoft has launched an artificial intelligence platform following successful pilots with telecoms, finance, energy and media firms.
- Indian IT services giant Infosys is using IPsoft’s technology to automate its IT and business process services.
- Accenture is using IPsoft automation technology to reduce the manual intervention required by customers in their IT infrastructure operations.
Such AI platforms will inevitably meet Luddite-like opposition as people protest about job losses. But Dube says this is only a short-term concern and should not hinder the adoption of these technologies. “Humans have a history of doing different and by freeing up the human mind, man could focus on other things, like colonising Mars,” he says.
For example, at one time, most people in the US worked in agriculture, but today only 2% do, says Dube.
He points out that only 10% of the human brain is normally being used and common chores take up much of this, so there is a lot of spare brain power to be focused on more important challenges.
Imagine, for instance, if Amelia could take over the management of legacy systems at big banks. This task consumes 80% of banks' IT budget and probably a larger proportion of their time. Amelia could read all instruction manuals and automated fixes and could possibly support legacy transformation.
Dube has big plans for his brainchild. “In 20 years, Amelia will have freed man from the shackles of common chores and allowed man to extend his horizons into creative dimensions,” he says. “For instance, Amelia would have taken over general physicians’ tasks, allowing doctors to focus on more profound issues, such as finding systemic cures for diseases like Alzheimer's and Parkinson’s diseases, rather than being caught up in day-to-day prescription writing.
“From a currently confused civilization, AI is the only way forward to systemically eradicate suffering and explore other worlds. History is witness that a civilization must innovate, or it will die slowly.”