It’s barely months since ChatGPT — an artificial intelligence (AI) chatbot — was released by American research lab Open AI in November 2022, and already many Indian companies are moving to adopt it.
Air India, for instance, announced it will use ChatGPT to modernise its digital systems, while tech major Tata Consultancy Services said it is working on similar generative AI tools — technology that can produce content including text, imagery, audio, and synthetic data — for enterprise solutions.
What makes ChatGPT so attractive to companies and how will it drive business?
Contentstack, a platform that enables businesses to create personalised content for audience engagement, has integrated ChatGPT into its ‘headless content management system (CMS)’, which helps businesses create content for delivery through multiple channels such as websites, mobile apps and wearables.
Also see:Why are investors betting on Indian AI startups?
With a few clicks, the AI assistant allows teams to quickly create summaries, outlines, metadata tags, descriptions, headlines, and even full-length keyword-optimised blogs.
Nishant Patel, founder and Chief Technology Officer, Contentstack, says ChatGPT allows teams to “create brand and tone-specific content in seconds”.
The one-click integration of ChatGPT into Contentstack’s headless CMS saves clients time and cost.
A coder’s dream?
A critical point of interest in ChatGPT and related generative AI technology is the gamut of capabilities it offers to those who code. While tech giants like Microsoft have integrated ChatGPT to allow users to develop applications with little or no coding, others are relying on the combination of human skill and the chatbot’s efficiency to elevate the coding process. This, in turn, has some developers worried about losing their job to a bot, while others are embracing the productivity it offers.
Also read: Elon Musk plans AI start-up, OpenAI rival
Data analytics firm Tredence uses ChatGPT on its AI-driven platform, ATOM.AI, to help clients and its own employees achieve productive coding.
Aravind Chandramouli, Head of AI-Centre of Excellence, Tredence, told Quantum, “We have employed prompt engineering on top of ChatGPT to ensure that developers, data engineers and data scientists can develop code efficiently. It enables them to seamlessly understand the code, check whether the code is efficient and make improvements. With this, we are already seeing productivity gains in our environment.”
Soumendra Mohanty, Chief Strategy Officer, Tredence, said that the company came up with ATOM.AI when it realised that AI solutions often involved problem statements (descriptions of research problems that can guide a company towards building the right product) that were so vague that it was difficult to break them down.
Entry of GPT-4
Some companies have adopted GPT-4, OpenAI’s latest multimodal large language model (LLM), which is a machine learning model that uses deep learning algorithms to process and generate human language. It can perform specific tasks such as language translation, summarisation and answering questions, among others.
Customer engagement platform Exotel, which recently launched ‘ExoMind’, a product powered by the AI and machine learning models of GPT-4, describes itself as ‘model-agnostic’.
Puru Govind, Chief Product Officer, Exotel, said, “We aren’t married to just GPT-4 or any other single model. We understand that our customers will have different scenarios and requirements and, consequently, we plan to work with other models such as Anthropic or LLama by Meta.”
Why Chennai unicorn Uniphore took over a French companyIIT-incubated Uniphore acquires Hexagone to strengthen its AI-led customer analytics
So, how does this work? “We are building connectors to various systems to ease the ingestion of data,” Govind explained. “We run LLMs on the data. Then, we expose applications in various forms such as a question-answering interface, or via bots, or APIs [application programming interface].”
The conversations created by an ExoMind bot, Govind noted, flow more naturally as they are more context-aware, leading to fewer misunderstandings for a better user-experience. These bots are also able to learn better on their own, allowing for continuous improvement without need for constant human intervention.
The bots can be deployed instantly, with better coverage of the questions that can be answered. Exotel also found that while it takes about a month to build a rule-based bot for a company, it takes less than 60 minutes to build it with ExoMind.
At the end of the day, hopping on the ChatGPT bandwagon early might fetch companies the big break they are looking for. But one thing remains clear: the ChatGPT/Gen AI wave is swelling, and it may be safer to surf it than risk going under.