As the hype cycle for generative AI and GPT-like technologies peaks globally, Tata Consultancy Services is also developing AI tools which code entire enterprise level solutions for their clientele.
In an interview with businessline, TCS COO, N Ganapathy Subramaniam, explained that TCS has been developing solutions that dabble with some form of automation for more than two decades. TCS’ cognitive automation product, MasterCraft, is able to reuse and rearchitect existing lines of code to build solutions. While the MasterCraft tool has been successfully used to generate code for specific applications, Subramaniam highlighted that TCS’ next steps are to adopt generative AI into the product and to develop a truly generative software that can build entire enterprise level solutions.
“We had always felt that there will come a time when we should be writing software that generates software. Our tool MasterCraft, is a repository tool that we developed two and a half decades ago, and it is fundamental to our well-being in many ways. For example, our banking tool, TCS Banks, is completely developed using MasterCraft. This means that not a single line of code is handwritten, everything is generated,” said Subramaniam.
Next step
While TCS has incorporated some level of AI into these solutions already, incorporating generative features will be the next step, according to Subramaniam. He believes that it can take six months to two years to build these tools.
Experts believe that generative AI technologies have the potential to create intricate algorithms that could solve enterprise-level problems. Therefore, the development of tools with generative features could spell trouble for many coders at TCS. The company is known to hire thousands of freshers, every year en masse; that have sufficient technical knowledge to write simple coding solutions at scale. With generative AI, such jobs could become obsolete.
Subramaniam muses, however, that this new technology could open roles for a new class of engineers that he terms “prompt engineers.” These prompt engineers will specialise in developing sharp, concise prompts for the AI bots, which will thus deliver the most accurate solution in the shortest amount of time on these inputs.
“An answer of a generative tool can differ a lot by the quality of questions you ask,” Subramaniam explains, “Thus a prompt engineer, will have a technical background, along with contextual knowledge of the business, client and a degree of design thinking,” he added.
Subramaniam believes that TCS will have to retrain its engineers to learn these tools. “The learning will be a lot more interactive and less bookish. We are still at the drawing board regarding what this training would look like, however, such AI technology will unlock immense potential for workers,” he said.