How Agentic AI will revolutionise personalisation

Personalised experiences are nothing new. From content recommendations and personalised ads to a bespoke user experience, video services are all trying to serve the user with what they are most likely to engage with.

However, personalisation must go beyond serving a user an experience that they will engage with at a point in time. Instead, it needs to dynamically change over time with the user’s tastes.

The challenge arises because this is usually very costly, and also risks not being exploratory enough to capture the rapidly evolving consumer preferences.

Serving a truly personalised experience that evolves with users over time means happy viewers who will consume more content and stick with the service.

However, we have not yet reached the holy grail of dynamic, granular personalisation on an individual level due to some key challenges.

Of course, the first thing you need for any level of personalisation is data, something that the M&E industry seems to have in abundance.

However, the data we have sits in silos across different parts of the video business, making it difficult to gain a unified view of the customer experience. Even when a 360-degree data layer is established, the challenge lies in turning that insight into timely and targeted action.

The reality is that if you really want to create a fully personalised experience, you need to understand in depth each aspect of the customer’s journey.

An OTT platform might have the application and infrastructure performance data and maybe even the business performance in terms of acquired users or engagement metrics. What you likely don’t have is all the things being handled outside of the OTT platform, such as customer care interactions, marketing performance, etc.

Getting that, up until now, has been too longwinded and laborious to even contemplate for any video business.

Agentic AI introduces the potential to make this process dynamic by orchestrating personalised experiences at an individual level. It uses a suite of AI agents, each with specific aims, that work together to create a unified management platform across all expertise, such as content, monetisation, technology, and customer care.

Crucially though, this isn’t prescriptive in the same way as traditional software. An AI agent might, for example, be tasked with monitoring the customer experience journey but you don’t need to prescribe how it comes to a decision on that or plans the outcome. Each agent as a standalone will only deliver on a specific mission, rather than your overall business outcome, but where it gets really interesting is the fact they can interact and work together to do just that.

The other crucial point here is introducing data from every customer interaction, not just what is happening within the OTT platform.

By doing that and using these agents, you can build up a much more detailed picture of each individual and use that to make their experience completely unique.

For Agentic AI to be really effective across the entire media lifecycle, we need to make sure that all software is contributing with its data and fitting a composable architecture.

Eventually, when everyone is using it, we can even do away with APIs as software will simply be able to talk to each other using these agents. However, if vendors are building Agentic AI solutions solely to enhance their own value proposition, rather than to also facilitate the cohesion of an ecosystem, then we will inadvertently be creating a new barrier and slowing down progress.

This is where we need to shape up as an industry to embrace the potential of Agentic AI.  

Bleuenn Le Goffic is VP business transformation at Accedo