Groww to launch in-house agentic AI model to help customers make better investment decisions

Groww, the country’s largest broking firm by active investors, did not respond to Moneycontrol's queries on the launch of the AI solution or the timeline for its availability.

The need for an AI solution stems from the fact that figuring out everything from an app user interface is still a daunting task for many users.

“We still need to go through a lot of steps. A lot of things are going to move from clicks and browse to more of a conversational kind of thing, and the broking industry also needs to do that,” the investor said, adding that going through a particular UI flow prescribed by a company’s design team is a thing of the past.

The Groww AI solution is enabling users to invest in their own way suitable to their profile, while the AI assistant can do it for them.

Horizontal and vertical AI models

Since the launch of ChatGPT, several investors have been taking screenshots of their portfolio or uploading on similar chatbots like Perplexity and Gemini, among others, to analyse their portfolio or to give them suggestions or recommendations for investment.

In fact, earlier in August, Perplexity expanded its finance vertical to cover the Indian equity markets, giving users access to live stock prices, news, and the latest information. The platform will also provide explanations of notable price movements and bull and bear cases across key stocks, along with access to company financials and historical data downloads, the company said.

However, such horizontal AI models lack the depth that a vertical and well-trained AI model can provide for their customers with deeper access to data.

“The data is never complete when you go to an external AI model, and they can get it wrong. There are several instances of that. These are horizontal products without any specialisation. An in-house agentic AI has access to all your data, investment patterns, risk profile and a vertical specialisation trained on platform data. So you get a better product,” said the second investor who accessed the Groww AI model.

Unlike many other platforms, Groww AI model lets its customers execute trades with a mandate and is not limited to strategy and analysis, investors said. For algorithmic trading, there is a specific coding, whereas through the AI model, it is a natural language command.

Along with the launch of Groww MCP (Model Context Protocol) earlier in August, the company is set to simplify a lot of use cases for the investors, while using natural language queries.

MCP is a set of rules and standards that allow AI assistants to talk to different software programs efficiently. MCP is supported by multiple AI platforms, including Claude, Windsurf, and Cursor. Anyone with a subscription to the Chat GPT/MCP server can build a solution with some coding.

The investors said that after feedback from the community in June-July, Groww decided to develop the AI solution that will help the customers who are not necessarily trained to use and integrate MCP with their Groww account.

For instance, to use Groww MCP, the users need a Claude Pro or similar subscription, which is expensive.

“AI is becoming our first kind of app. Most investors are uploading their data and screenshots on ChatGPT or Perplexity to analyse and ask for suggestions from these AI chatbots. It would make more sense if the company developed something in-house based on these AI agents,” one of the investors said, adding that the idea is to enable them to do more in less time.

However, while giving these data sets to the outside AI chatbots, the investors end up giving a lot of context which is not required for an in-house solution, as it has access to all the data.

According to the investors, the idea behind the AI solution is that it should feel like another feature within the app interface, and be easier to summon because of the natural language and the lack of a particular flow, unlike other features.

Move over MCP, brokers look at in-house AI solution

It is not just Groww that is bringing AI models to the customers. On August 26, Moneycontrol reported that Raise Financial Services, the parent company of stockbroking platform Dhan, is set to unveil 'Fuzz', an agentic artificial intelligence (AI) model engineered to synthesise deep research, regulatory filings and official financial data for investors.

Bengaluru-based FYERS, which has 8.5 lakh customers, launched FIA in April this year to let traders bypass the traditional filters and interact with market data through simple natural language queries.

The investors said that MCP does not work for the large segment of customers, as most of them do not have access to subscriptions or are not well-versed in the integration of APIs of different systems.

Earlier in May, Zerodha forayed into AI-assisted investing through the brokerage firm’s new MCP-based portfolio management for more accessible and AI-backed decision-making.

“AI tools have become so good that you don’t need a UI anymore,” said Zerodha CEO Nithin Kamath in a LinkedIn post showcasing user response.

Interestingly, there are no clear regulatory norms on the use of AI as an assistant or investment advisor as yet.