Intelligent decision automation tool from TCS

CIOL Bureau
New Update

PUNE: Tata Research Development and Design Centre (TRDDC), the research arm of TCS, developed a memory-based decision automation engine, the TCS Consult.


TCS Consult helps solve problems of a repetitive nature like customers' queries, approving loan applications, processing insurance policies, etc.

Over 50 years of R&D person years have gone into the development of TCS Consult that can operate in two modes — Reuse and Learning.

In the Reuse mode, whenever a problem arises, the engine conducts a search for previously solved similar problems. If such problems and solutions exist, it suggests the solution used earlier.


In the Learning mode, if no prior solution exists, the engine acts like a listener, watching the experts at their job and learns by associating new problems with the expert solution. Next time a similar problem occurs, the system will solve it.

According to Vivek Balaraman, research scientist who heads the Decision Systems Initiative at TRDDC, over 95 per cent of problems in the industry are of similar nature. In such cases, the solutions that worked earlier will work again. Therefore, it is possible for solutions to be reused again.”

Mathai Joseph, executive vice president, TCS and executive director, TRDDC said the tool helps in quickly understanding the client's applications and maintains high solution quality across shifts and geographies. He cited examples of a US financial investments house that spent over six hours over a problem. After introducing the tool, the solution for this problem took just an hour.

The tool also finds applications in business scenarios where the issue involves mining knowledge buried in text. A US software company faced the prospect of supporting costs of $1 billion per year in solving transcripts buried in emails and logs. The goal was to reduce costs by reducing incident volumes. The TCS tool here helped reduce costs by reducing incident volumes.

The first version of the tool is to be released in November. Joseph said it would first be tried out on mid-sized problems.