Transcript
esynergy is a tech consulting company focusing on delivering value to highly regulated industries like banks, insurance and public services. Predominantly in the area of data, AI and cloud. For our customers ranging highly regulated industries, it's important for us to understand what problem they need to solve and AI is a tool to help them in solving those particular problems.
Initially when we are building this sales AI assistant we could have done chucking all these unstructured documents using some OCR or some. But that is not helping our purpose. Our purpose was on looking at specific keywords what was required for us to retrieve and augment it with a language model.
And Vector is a great store where in which it can help in creating right kind of embeddings and faster search and bring right kind of context so that I get what is relevant for my context and I get what is relevant for my customer.
And sales can look at specifically for example searching for have we done any Kubernetes related work in cloud migration? One of the key you know, benefits that we got it from our AI sales copilot the time in which the sales go around and spend time in finding the right case study, right kind of experience, right kind of work we have done for the previous customers.
Instead of taking weeks, they were used to doing hours. So that's a huge advantage of one making it real and in the context and getting it faster and which is appropriate for a particular customer. Narration why we chose Astra DB was on its ability to scale and it is too serverless and cost management was very critical for us having a cost management effectively and making it available as part of, you know, so most of our engines were built around AWS.
So having an AWS Bedrock as a as a key entry point for the AI solutions and Vector as an Astra was giving us a perfect combination of, you know, getting scalability as well as making it sure that we have right kind of cost management around it.
For anyone who is starting the journey in building gen AI apps, it's easy to start a POC very quickly today because most of the services what you need is available on cloud whether it is in AWS or it is in Azure.
So you only need very few services to start with. So as a developer or as someone who's trying a GenAI app, the barrier to entry and getting the right kind of environments is. Is much much easy now right we can get DataStax Astra as a service we have Bedrock will give you multiple options of choices of language model.
We got LangChain to orchestrate and stitch these things together in few days you should be able to get initial results of a POC and GeniAI apps. Always you have to experiment, learn and continuously improve, right? So those are things which it is an important thing for anyone who getting starting in this journey.