Unified intelligence: Your operational decision layer.

An ongoing challenge here at Entopy, and one I’m particularly responsible for, is product messaging and positioning. Are we a Digital Twin? Are we AI? Are we data orchestration? We’ve experimented with all of these and tried different combinations and sequences. But nothing has quite hit the exact remit of what we do. Digital Twins mean many different things to many different people. AI on its own is wayyyy too general. Data orchestration misses the intelligence. 

We’ve now landed on Unified Intelligence. This is our working attempt to capture the many facets that Entopy’s technology focuses across but critically, plays directly into the ‘so, what?’, and that is ‘to deliver the operational decision layer to complex operational environments or ecosystems to support operators with high-consequence, highly complex decision making. 

What is it? What specifically do I mean by Unified Intelligence?

At our core, our mission/purpose is to help operators in complex environments make better decisions. We do this through a composition of technologies which includes data orchestration, AI (and the various profiles of that), Digital Twin (more specifically, things like ontologies) and of course data. 

Over many years, we have developed a platform that delivers advanced intelligence specifically in the operational context. We start by bringing data together from many systems. We then deploy targeted AI Micromodels to specific parts of an operation to predict specific aspects. This type of AI is good old-fashioned machine-learning. We then integrate the data and outputs from Micromodels to form intelligence workflows which deliver insights. These insights can then be displayed through dashboards, pushed through channels like MS Teams and are available to our AI Agent (the very much in fashion LLM). The LLM uses the underlying stack of integrated data and ML models to deliver powerful intelligence through a chatbot interface, via situational reports distributed by email or within the software recommending actions in real-time, illustrating cascading impacts and so on. 

Why so complicated? 

As you can see, it’s a powerful stack that comprises a myriad of technologies. And to deliver effective intelligence in the context of operational decision-making, you need the full stack. The key with operational decision-making is that it is constant. Operators are constantly assessing the environment/ecosystem for possible constraints or issues to get ahead of them. This is to a degree different to other decisions that may occur within an organisation that maybe have a shorter decision-horizon (more about that in another blog). 

Operators need dynamic, real-time models that evolve with the complexities as they unfold as opposed to static models that are typically used for more static or longer horizon decisions (hence real-time data integration). The intelligence needs to be auditable (hence granular AI Micromodels). If I’m going to change a specific aspect of my operation based on a prediction, I need to understand and validate that model (and honestly, I want to see it). Complex operational environments often involve many stakeholders and for effective intelligence, this data needs to be brought together in a way that protects the underlying security/privacy of the respective holders (hence ontology/orchestration/segmentation). I want the intelligence delivered where I already work (hence notifications & sitreps). And I don’t just want an alert; I want suggestions about how I can remedy it (hence LLMs). 

A new category.

The myth that one can just plug in an LLM and away you go has dissolved. With enterprise adoption still <5% and many projects stuck in the pilot phase, something is clearly wrong. Not with the vision, but with the method. An LLM with a bit of data and context is cool. An LLM with a web of intelligence through a composite stack of data and intelligence, is an operational decision layer.  We believe this category will define the ways of working for many industries over the coming years. 

Written by Toby Mills, CEO.