AI Assistant

Hot on the heels of the improvements mentioned last month with regards to our unified data model, we have now released a beta version of our AI assistant. The assistant enables users to ask free text questions of their data without the need for a query builder or SQL knowledge. Answers can then be visualised as a chart or table, so a user can go from asking a simple question to defining valuable insight in minutes. Read on for specific examples of how to use the assistant.

Global Assistant

The Data Refinery now has a global assistant that can be accessed at any point to ask questions of your data held in The Data Refinery. Use this as a starting point for understanding key metrics or for suggested improvements or actions that could be taken based on historic data.

The answers provided can then be visualised and if valuable, charts can be added to existing dashboards.

Dashboard Assistant

When building a dashboard or specific chart, the AI Assistant can now be used to generate a chart directly from a plain text question. This takes care of all the steps typically required to build a chart within The Data Refinery, the filters and metrics used to create a chart can then be inspected or tweaked to fine tune the visualisation generated.

Both of these features are in Beta and we will be monitoring usage to ensure we return the very best responses to our customers questions.

Subscription Data Models

Our unified data model now supports subscription or membership data types, this means data from loyalty applications or subscription-based services can now be integrated and harmonised with all other data models. We will soon be releasing automated reports and insight that is based upon this new data type.

Propensity Models

With our recent release of machine learning powered forecasts (see last month’s roundup for more information). We are happy to announce the release of Propensity models for customer entities. This is an initial roll out of a percentage marker for customers that will indicate:

• The likelihood of a customer re-purchasing
• The likelihood of a customer churning

These propensity figures can now be used to validate potential actions or segment analysis and help to ensure the most appropriate action is taken when communicating with customers, be it a gentle nudge for customers that are likely to re-purchase or a more personalised approach to customers that might be sticking around.

Events and Conferences

The team have been incredibly busy attending and exhibiting at multiple data and retail focused events throughout September.

Big Data London

We descended on mass to Big Data London and had some really interesting meetings and conversations with customers and suppliers.  

Key take aways from the event were:

• The modern data stack is becoming commonplace, luckily The Data Refinery is built on such a stack.
• AI and generative AI were the buzz words of the day, our latest AI features stand up well to the offerings at the event.
• Our partners, Snowflake and Fivetran, were centre stage, cementing confidence that we have the right tools in place.
• We have a blog post dedicated to Big Data London, read all about it here

E-Comm Expo

A subset of the team returned to London again a week later, talking to retailers and e-commerce teams about the value of data and how to transform business strategy using data that is housed across a typical Sale/Marketing tech stack.

The two days were incredibly busy, and we were blown away by the interest show in our product. We will be back next year with a lot more to say.

General Improvements

See below for a list of minor improvements and bug-fixes that have landed since our last roundup:

• Resolved an issue where sorting of dashboard data was sometimes not working.
• Resolved an issue where chart templates did not save correctly.
• Resolved an issue where in some scenarios the data search function would fail.
• Resolved an issue where in some scenarios invoice totals were missing from the billing screen.
• Resolved an issue where an incorrect cache is used depending on the order of screen views.
• Updated several backend services to improve performance.
• Added a new connection authentication link to allow users to delegate connection setup.
• Updated the primary navigation to include a dedicated Customer Menu.

‍Roadmap

Our roadmap can be viewed  here, and is ever evolving, we will ensure that all future features are based upon and prioritised via customer and community feedback.

Want to see these changes in action? Drop us a message and we'll show you how The Data Refinery can work for you! Alternatively, sign up and try out The Data Refinery for yourself.