With the advent of ChatGPT, the significance of large language models has been realised. These models permit to move from syntax to semantics, and improve our understanding of users.

In searches, there was just a syntax forward approach using certain key words. However, with the access to LLMs, the approach is to use semantics where the model understands the query without any reference to the key words. Human beings understand the surroundings this way.

The concept is being used in advertising. LLMs understand what the users want when they type something. They pull up the most relevant ads. There is more likelihood of users clicking these ads, and the advertisers gets return on investment.

The processing speed is remarkable. It is on account of specialised hardware. That makes the models to get trained faster.

Google uses LLMs in advertising. It is involved in building ML models, especially on the conversion side — ads that generate a customer response. Conversions improve by a significant 13 per cent on an average when automation tools built on LLMs are used.