One of the most exciting moments for our team is when we are able to put new product capabilities in the hands of our customers and partners – capabilities which offer the power to enhance the digital experience that they provide to their banking customers. Today, we are excited to announce the release of KAI 4.0. This major release provides crucial capabilities to our customers and partners in the mid-market and enterprise segments.
This release continues our efforts in offering the financial services industry not only the best conversational AI technology, but equally as important, offering a platform that is flexible, simple, extensible and economical. KAI 4.0 brings our customers and partners the empowerment to imagine, develop and design their digital experiences, while leveraging proven conversational AI capabilities and features, that are designed and pressure tested by an amazing and talented group of engineers, data scientists and product experts here at Kasisto.
Here’s what you can look forward to in KAI 4.0:
Multi Tenancy
In KAI 4.0, our industry-leading conversational AI platform will also be available on a multi-tenant infrastructure. Our customers in the mid-market and enterprise segments, can both benefit from this new architecture.
Enterprise: global financial institutions often provide their services across many geographical locations. Those countries may speak the same language, but the locale often differs. With the multi-tenancy, customers will be able to deploy a single NLU (natural language understanding) model across all countries, while giving locale customization capabilities to each country (e.g. differ on dollar vs. pound, neighbor vs. neighbour, and 12/25/2020 vs. 25/12/2020). This reduces the overhead of managing the virtual assistant significantly for customers.
Mid-Market: national and regional financial institutions often are more price sensitive, compared to their global counterparts. By running our workloads on a multi-tenant infrastructure, we are able to offer KAI to mid-market customers, so they can offer a rich digital experience to their customers. Many of our customers are regional banks in the United States, which often leverage a digital banking platform provided by one of our partners – those customers will now be able to provide conversational AI capabilities to end users at a lower cost, using one of our partner platforms.
Publishing Workflow
Building, managing, and deploying a virtual assistant requires contribution from a cross-functional team. From intents to entities, responses to segments, training data to models – there’s a lot to manage. It is common within an organization to bring together folks across teams and departments to work concurrently on the virtual assistant. The concurrent contribution has historically risked lost progress and overwriting.
Kasisto’s new Publishing Workflow simplifies the virtual assistant development process by giving teams transparency into updated content, while providing the required checks and balances to ensure the right content is being deployed to their end users. The Publishing Workflow provides easy access to version control, on-demand audit trails, maker/checker quality controls and multi-target publishing capabilities, enterprises can now ensure that the appropriate content is deployed to the proper environment at the right time.
Intent Discovery
Our customers see millions of users interacting with their virtual assistant on a monthly basis, and tens of millions of user requests through those interactions. While KAI is able to understand the majority of the user requests, there are always new questions that users want answers to from a virtual assistant.
The existing process requires a human team to manually go through the user requests individually, and add an Intent label to those user requests – as you can imagine, this process is tedious, expensive, and time consuming.
In KAI 4.0, the Intent Discovery (Beta) feature uses advanced machine learning techniques to review all user requests to (a) automatically group together user requests that are semantically similar (b) match the group to an existing Intent that the virtual assistant currently understands, along with a confidence score. Through this feature, customers will be able to automatically identify new Intents that should be added to the virtual assistant, or add new training data to existing Intents, thereby reducing the manual work significantly.
Kasisto is committed to continuing to lead the financial services industry into a new era of conversational AI banking by developing and delivering the industry’s most transformative conversational AI experience through an extensible and economical platform.
If you are a Financial institution or provider looking to step into the new era of banking, let’s talk.
