Rezolve Ai Investor Day: Commerce Infrastructure Technology - Sauvik Banerjee - REZOLVE AI PLC

Good morning, good morning everyone. My name is Shovik Banerjee. I'm the group CTO for Resolve AI. Having built large tech businesses, having led large technology organizations from SAP, Accenture Digital, and of course having met Dan two decades back in our journey to building vendor, e-commerce and AI is something which has kind of become the DNA of my existence. Now imagine the world we are living in today. Some of us have built these e-Commerce engines over the last few years and you saw how unparallelly we are chartering this territory agents are going to swarm across you've heard Lawrence you've heard David you've hurt Michelle now today I'm going to talk about what's the core technology which differentiates us this is what you've been hearing from the morning get chosen the discovery part get paid the checkout part get loyalty James just spoke about it and the whole flywheel of data comes back get chosen again this is the commerce this is the this is the future of purchase what we are talking over here is that's the stack but what runs that stack that runs what it runs the stack is brain power which is our models our voice orchestration our agents our agent orchestration and of course underneath is the ethics and compliance stack and of course we've got the whole database as a service which is which is something very exciting we are building on i'm going to talk about in a few minutes that's that's the technology backbone of what everyone has been explaining from the morning i want you guys to take 10 seconds on this slide to understand these three layers actually runs the entire front-facing engagements and is ready for a retailer to take on the swarm of agents and accordingly let me just delve a little deep on brainpower I'm going to talk about the model part of brainpower today though the orchestration the voice the agentic entire engine is all brain power in the day in October 2026 we keep hearing the world's large language models is where the transaction will happen we keep hearing in inaccuracy now if I look at what you see over here is a live commerce environment imagine it is expensive the real world problem is these purpose built LLMs which you guys use all the time is is trying to guess now think about the jacket use case Michelle was talking about machine has to know before it can self now none of that has ever appeared on the internet this catalog which is getting trained on brainpower is is trained on the catalog in itself it is not trained on internet words now when when we when we build this this not didn't get built with the explosion of LLMs in the last 24 months this got built over the last 10 years and it got built on single industry category 300 billion parameters 30 billion parameters 300 billion tokens and more importantly it's the results proprietary models what we do with when we build it ourselves is we can't be the maker and checker we have to then benchmark it against everything else for exists and that's that's the part where the catalog requires the insight the model can rank model can pattern match model can converse extraction of the attribute comes across the mindset of the sentiment comes across but it still cannot do it cannot hallucinate it cannot model drift because we are playing with retailers brand it cannot have an impact on the erosion so let me tell you what we do and how we define that this is this is the brain power resolves ethics ethics core we've got trace where now a common failure with AI agents is that it reports actions but which never took place example imagine an agent agent is looking at a screen it It will go through the screen. It doesn't understand the order management. It doesn't a transaction. It might report an incorrect transaction which never got placed. Now what Traceware does, it through logs, it rebuilds that journey and says that transaction happened and that transaction didn't happen. So you have a maker and a checker framework. Our trace where our model runs between 99.5 to 100% accuracy on these sort of logs we check on the agents journeys it's important because you can build a model but you need to check it right and then comes auditable AI you've heard us talking about hallucination free and model drift but we need that to benchmark it against so what we did was the platform of auditable AI gifts the outputs against other language model outputs and that's extremely important for us to be neutrally judged from anybody who is commercially engaged with us it gives the brands and the users a complete observability framework a complete monitoring framework in what it intends to achieve and AI behavior is something we are dealing with it's human like but it's not a human being so we got to be very very careful in what we are deploying to these platforms across these across geographies so we have to we are very very focused on what that does now we've been in the business for quite some time and something which we had built now has just taken off is any output of a catalog image video audio of a PDP we have a technology today which actually does the verification that's resolved provenance resolved provenances exactly what it is it embeds invisible watermark on an image or a video or an audio and a text the future of PDP just got transformed and more importantly if I take it offline it can work on multiple form factors so we're very very deep tech company and and and this has been my founder chairman CEOs vision this has been the vision of my chief scientist and chief AI officer dr. Salman Ahmed and some of these brilliant minds i get the pleasure and opportunity to work with now let me show you what's the most exciting part we are working towards you heard naga you heard all our stacks on ai now imagine this we grew up in a world of columnar and road-based databases then came the world of in memory you know the Cassandra's the Hannah's of the world today which is the network which is infinite scales in multi-milliseconds is the blockchain imagine a catalog 20 million catalog imagine a hundred million day of transaction like you know festive days we used to hear and I used to grew up with database chokes systems would be down no this is where the world is moving forward the brain powers database as a service powers the future of catalog the future of e-commerce transactions the future of concurrency on a user on a cart we disclosed it in our annual report Dan mentioned it that we're going to make it commercial it's almost there we are going to now roll it out across across our stacks which you saw is where we are championing database distributed database as a service and we're always inventing we're inventing, we're pioneering that's our DNA Resolve AI labs is one of our research divisions we've got our teams across the globe right from here in the US across various places in Europe it is there in South America and of course it's in Asia but I run the team in India being the group CTO as well and and every day working for resolve is building inventing and what about my team calls all of us is problem-solving utopians so that's the world we are building thank you for listening in I'm going to hand it over to Elizabeth and Steve Perry and where we are going to have an engaging conversation over to