Extract value with assisted analytics

Key Takeaways

[Music] welcome to next24 here we are demoing how Gemini can help you with assisted analytics right so Gemini has a new feature embedded now called function calling what that fun what allows you to do is make your SQL developers and the complexity of developing SQL kind of minimize right what it also allows to do is any business users who are not SQL C interact with with the backend systems and the systems could be anything could be relational database elix database CRM Salesforce systems to run and run queries in a non-bus fashion right that's what it does so if you look at the challenges that it solves in the past when you are dealing with these large language models often times those models are frozen in time in a point in time what that does is makes your information outdated right so let's say for example if you're looking at a currency exchange rate and if you look at the query that je the L May say well I don't have the data as of today but I can show you something from last year but that's not what you want you want realtime exchange rates so it helps you solve that it also helps you embed and pull data from multiple sources real time so it's not a static data even though it's grounded to certain data elements but you can also invoke certain ex external API calls to push the data in and it also gives you a very consistent view of the data as it exists so in sales order you need to make sure that the sales order the sum of the the sales amount is right on the money right so it does that so what Gemini does is really makes your complex workloads much simpler right but providing that Automation in the back end that's what the power of the Gemini is and it also allows you to do enable some workflow automation so if you're in a help desk and you want to kind of have a chatbot like experience it gives you that capability as well so why don't we show you the demo real quick and Kenny is going to now show you the demo of this product all right thank you man now let's us switch this is the small application we build this is the python app they only have a 270 lines of code you know you see here we have a sample prompts there unstructured written natural language any of your business analyst or Executives can access database without knowing or without any skills of SQL uh query language let's say pretend I want to know and in my database hey what a percentage of all the return okay I just to interroga my database with that kind question the c c with a bigquery that goes on they make it a f function calling then rain the result so the right answer come back and say 99.95% order are return how do we get that conclusion here is the SLE quy we generated and you see in the whole process let me show you more details here you can see this is the function call we make we list what database within the database how many tables we have seven tables then within the table say hey where is the order order is here we point to the order table then we generated the SLE query this is sqle query return that 7% almost the numbers all right let's see what's the source code behind this okay you see that's all 279 of code in this python code we did not tell you we don't tell you how to write a CLE all the Gemini pro model smart enough they do automatic translation they get your natural language question convert and generate a function code and based on that function then we do do Loop until we answer we find the final result for the question you were asked let me turn this back to my teammate man and he can explain that next one so just to summarize what we showed you is how a a a non-technical user can interface with Gemini and Gemini will do all the coding on the back end right so what it does is as a user I simply provided a prompt and it the prompt will call the right functions which runs the SQL query against the backend system right in this case we're showing how it's going to pull data from bigquery one a Google analytical platform and then it provides bigquery provides a response back to the function and that function then transfers the response back into a human readable natural language and that's what the that's what the user is presented so as you can see you start with a very basic English language the the the Gemini takes care of all the technicalities and then you get the answer back in a plain English there are multiple use cases that Gemini can work with we demonstrated one use case but if you have Finance systems databases documents right unstructured data you have support systems like chatbot that you want to build it can allow you to interact with all kinds of systems as long as it has an open API call that Gemini can make to it and so as a call to action you know we have lot of stuff going on the next 24 it's a very popular event we would like to have folks look at this QR code you can get the source code for this that we demonstrated you have very good articles published in the blogs that they can read up more about the function calling which is a new capability launched just two months ago in Gemini and you can also get some sample code development tools through this URL link as well thank you so much for your time come to play at a home go for it