We take a value approach to AI
We constantly innovate at Epi and embrace the opportunities and challenges that artificial intelligence (AI) brings with both open eyes and an open mind. The rise of AI is highly complementary to our mission to connect data, science and people with full-stack open-source data science and software engineering.
This is why we – where it adds value – enhance our dashboards and other data products with focused AI.
We do this in two different ways.
Firstly, by what we have always done – integrating complex models into Shiny apps - predictive machine learning models in this case.
Secondly, we integrate external AI. Take, for example, the dashboard we built to support monitoring of biodiversity in Australia’s Marine Parks. It starts with a traditional layout to explore fish species across location, but it also lets users “chat with the data” via an embedded AI-powered assistant. The dashboard sets the context; the Large Language Model (LLM) unlocks deeper discovery and can add written summaries that explain what’s happening in the data. In the Epidemix app, which bridges the gap between science and policy by making disease models more accessible, users can send a prompt to an AI chatbot to get assistance with selecting parameters for a specific disease of interest.
This hybrid approach enhances our dashboards, by making them more interactive, yet the users are still guided through the analytical process by tailored information design and provided with the strategic context for their queries.
But won’t AI queries eventually make dashboards redundant?
An AI chat bot (or agent) might be more suitable when a user needs a highly specific and immediate answer on an ad-hoc basis. But will AI completely replace dashboards? We don’t think so! Dashboards, reports and pre-defined data products simply serve a different purpose – they provide essential structure, context, and repeatability. Especially when monitoring change, supporting decision-making, or aligning stakeholders around a shared view. Importantly, dashboards are deterministic: no matter how many times you repeat a query, you always get the same result.
However you choose to analyse and work with your data our team is here to provide support. Our full-stack approach enables us to address your frustrations and bottlenecks, wherever their origin, and embrace new technologies like AI with open eyes.
