Julia Bondarchik
Head of Product
icon

YOKOT.AI – lessons learned from our customers after 4 months in production

It’s been 4 months since we have launched YOKOT.AI – our private generative AI solution for business use. Since the launch we have run pilots with over 20 different customer companies and partners, onboarded 100+ users and learned many valuable lessons along the way. Generative AI tools and applications are still relatively new to the common public, and we often see a lot of confusion around the capabilities of such tools and how to get the best out of them. That is why our Softlandia team not just offers you the tools, but also provides guidance and support along the way.

In this blog post we are discussing the importance of finding and defining just the right use case for your AI solution, why it is more efficient to utilize a complete RAG architecture instead of figuring out your own solution and how to ensure your corporate data remains private.

Define your use case and the process

The most common expectation we see from the customers is that generative AI is a magic box, where you can throw in all of your knowledge and expect it to read your mind at the press of the button. Unfortunately that is not the case, or at least not quite yet 🙂. In order to get reliable answers from your data, you first need to feed the AI good quality information relevant to your use case. The magic starts once you can clearly define and describe your expected outcome in a similar way you would need to explain it to another human being.

We are well aware that an evolution of a work process can be time consuming and a hefty task, but you don’t have to do it all by yourself! During the user onboarding phase, we offer our expertise and training to get you started. Along the way, our team of professionals is always ready to help you to set up the tools, figure out your use cases and find the best way to imbed AI tools into the processes of your organization.

Take into use a ready solution, rather than figuring it out from scratch

We have built a complete software architecture employing modern AI technologies and approaches, that can be deployed to our customer within 24h and ready to use. Our sophisticated RAG framework at the core of YOKOT.AI allows you to customize and tailor the solution according to your unique needs and objectives. By feeding your data into YOKOT.AI database, you are able to create your very own knowledge base no stand alone off-the-shelf language model can offer.

RAG architecture is implemented in Softlandia’s private Azure environment

In addition to getting insights from your documents and data through the chat interface, we have planned a lot of cool features for YOKOT.AI to be rolled out early next year. First automated workflow has already been tested by our customers and gained great feedback! Document Generation workflow is able to create a complete document draft defined by your needs and derived from your own sources.

Satisfied customer feedback

However, not all use cases can not be currently executed with YOKOT.AI web application alone, that is why we offer the YOKOT.AI API that can be integrated into your existing services and tools. One common example we came across is retrieving data from structured databases with semantic search powered by language models. Though, you can not (and probably shouldn’t) upload your entire database into YOKOT.AI data management system, it is possible to deploy YOKOT.AI API on top of your existing database to supercharge your data queries.

Make sure your private data remains private

During the past year, we saw numerous cases in the media of private data leakages and a number of examples on “how to trick chatbots to reveal the data”. This is why you should be extra careful when selecting your technology provider, especially when working with organizational internal knowledge and confident information. It takes a lot of background work to ensure that your data remains private and stays within your organization by incorporating various guardrails and security measures.

Our customers highly value their data security and privacy and we can assure you that unlike public alternatives where data may be processed on external servers, YOKOT.AI keeps your sensitive information within the confines of a GDPR-compliant environment.

If you have been waiting for the right time to start exploring the AI capabilities and applications for your business, wait no more! Contact us now to enhance your productivity in the year 2024, isn’t that what New Year’s resolutions are all about? 🙂

In the meantime, we would like to thank our amazing customers and partners for a great year and all the lessons they’ve helped us to learn so far! We wish you all happy holidays and look forward to continuing discovering the rapidly evolving scene of generative AI together!