Augmetrics®

Accurate, consistent, cost optimizing and energy-efficient.

There are many things to know about Augmetrics®, but they can be wrapped into a single definition:

Augmetrics® is a new form of energy efficient, highly accurate data processing that generates permanent information objects that are processed once, updated according to your rules, remembered forever and are available to any permitted user or use.

And one more thing…

While we believe in the power of AI, fundamentally we believe in empowering human experts, not replacing them.

We do that by training smart machines to provide them the information they need to help them make expert decisions – Leveraged Human Intelligence®

Our
Applications

Chloe NLP™

At the application level, your empowerment partner is Chloe NLP™ , a natural language analytics agent whose job is to bring you – any user – the information you request together with an interactive language equation (ILE) showing you exactly where that answer came from and how it was calculated.

In Right Answer Mode, Chloe’s answer accuracy is hard to and as a natural language agent who minimizes LLM interactions for market leading cost control capabilities.

Reportworthy™

Reportworthy™ automates and delivers all your data-informed recurring and narrative reporting, saving 1,000s of hours of formatting, set-up, analysis and writing per year.

The Path to 100% Accuracy

(at the lowest sustainable cost)

Phrase variant creation, audit and caching.
For each successful query resulting in an information object, Chloe submits a prompt to generate up to 100 variants (basically synonyms) and then a second, slow-roll prompt to audit them. All successfully audited variants are stored in the query layer cache.

Default tokens (information objects).
For each data source, Augmetrics® pre-builds hundreds, even thousands of information objects. The net result is users start out with every basic combination of analytics dimensions, metrics and common date intervals at their immediate fingertips. The queries to create these objects are then run through the variant creation steps and you get a per data source vocabulary of tens of thousands of proven natural language phrases that will generate successful queries.

ILE, the interactive language equation.
You probably remember the natural language equation that Chloe returns with the answer. This week the highlighted words in the LE will be interactive so the user can click on them and find every applicable data field that can replace the bolded dimension, metric, attribute or interval (date). So if the LLM gets it wrong or the user just wants to noodle around in the data, they can change any of those fields and Chloe reprocesses the query.
In Right Answer Mode, all of these helpers drive toward CPUs, not GPUs…meaning the less the customers interact with the LLMs, the faster they have the path to 100% accuracy.

And…the cheaper the cost per query!

“In a world that increasingly boils oceans of data for minimal gain, the risk of boiling the planet exponentially accelerates as those models fail to deliver on informational speed to value.”