The important decisions are often made in half-light.
Clear thinking about data and AI under uncertainty.
Why Penumbra
Penumbra Intelligence takes its name from the penumbra: the partially illuminated region at the edge of a shadow, between full light and full dark. It is the space where boundaries blur and understanding is incomplete, and where insight is most needed.
In fields being reshaped by data and applied AI, the most important decisions rarely sit in clear light or total darkness, but in these transitional zones between data and judgement, theory and practice, possibility and constraint.
Penumbra Intelligence works in this space: helping organisations interpret uncertainty, integrate emerging technologies thoughtfully, and make robust decisions where signals are noisy and systems are interdependent. The name reflects a focus on clarity without oversimplification, and on progress grounded in careful reasoning rather than hype.
What I do
Most organisations do not lack data. They lack a clear view of what it can and cannot tell them, and a way to decide while the picture is still incomplete. That is the work.
Advisory
- Data audit
- What data you actually have, what state it is in, and what it could support.
- Data interpretation and decision framing
- Making sense of your own data, and structuring the decision it feeds.
- Technology fit
- Vendor-neutral advice on which data and AI tools suit your needs, and which you can skip.
- A second opinion
- A short, bounded sounding-board session on a decision already in motion.
Education
- Data and AI literacy
- Briefings and workshops that give a team a working understanding of what data and AI can and cannot do for them.
- AI myth versus reality
- A briefing that separates what AI can do today from what it is claimed to do, so decisions rest on the first.
- Presentations, workshops and talks
- Keynotes, invited talks and in-house sessions on data, AI and decision-making.
How I work
- Independent. I sell no software and take no referral fees. There is no vendor I need you to choose.
- Plain language. If it cannot be explained to the person who has to live with the decision, it is not finished.
- Honest about limits. If the data cannot answer the question, I say so, and say what could.
- What happens when you get in touch. A 20-minute conversation to understand the decision you are facing. Then a short written scope. Then you decide.
About
I'm Jordan Pennells, a chemical engineer based in Melbourne. My doctoral research at the University of Queensland developed a statistical framework for evaluating the quality and sustainability of plant nanofibre materials made from sorghum crop by-products. Underneath the chemistry was a problem I have met in every job since: how to judge and predict an outcome when the raw material, and the data, never arrive the same way twice. I now work as a research scientist at CSIRO, applying data and AI to processing problems in food and agricultural research.
Much of the work I enjoy most happens in front of people who have to make decisions about data and AI without being specialists in either: industry audiences, webinars, boardrooms. Penumbra Intelligence grew out of those conversations, as an independent practice alongside my research role. I also build software and models myself, which keeps the advice grounded in what is actually feasible.