Evidence before certainty
Facts, assumptions and unknowns are kept distinct. A source does not become reliable merely because it is available.
Decision support framework
Better decisions begin with knowing what you actually know.
Reports, emails and site conversations do not always tell the same story. Yet a decision still needs to be made.
Rabbit Opt uses AI to organise the available information, identify gaps and conflicting accounts, and prepare options with a recommendation. You can examine what supports the recommendation and what remains uncertain before making your decision.
We are currently testing the pilot in construction.
A fictional pilot example
Can the work proceed?
The planned facade closure is approaching, but the records show an unresolved QA hold and an unverified condition that closing the facade would conceal.
In this fictional demonstration, Rabbit recommended deferring closure and identified the checks and formal release needed before reconsidering it. It also kept the unknown time and cost implications visible.
Read the case summary and full report
Why I started Rabbit
Tomasz Szczęsny
I’m Tomasz Szczęsny. I started in UK construction as a labourer, progressed through carpentry and site supervision, and went on to run projects. Later, studying Construction Management helped me connect practical experience with the theory behind decisions.
On site, I saw how much still relied on paper notes and conversations. A laptop was not always practical where the work was happening, but a smartphone was already in my pocket. That raised a question: could it help us capture information and make better-informed decisions?
That question led to Rabbit Opt. I’m now looking for construction practitioners to test the pilot and tell me where it helps and where it falls short.
Tomasz Szczęsny on LinkedInAn invitation to test
Try Rabbit with an example or your own case.
You can start with a prepared fictional case or describe your own construction situation, without confidential or personal information.
I’m looking for honest feedback: does the analysis help, what is missing, and what makes the pilot difficult to use?
To request access, email rabbit@rabbitopt.com with a few words about your experience.
The decision flow
Evidence → What is known and uncertain → Risks and options → Recommendation → Human decision
The framework can also record what happens after a decision, so its basis and outcome can be examined later. The current Web Pilot is a controlled review build; this page describes the wider framework where appropriate.
From the Rabbit philosophy
Purpose first. Evidence before certainty.
Rabbit begins by asking what the decision is meant to achieve. It keeps uncertainty visible and seeks only the additional information needed for a responsible recommendation. The depth of analysis should match the stakes; the authorised person remains responsible for the decision.
How Rabbit thinks
It does not tell you what to decide. It shows you what you are deciding with.
Conflicts stay visible
When sources disagree, the disagreement becomes part of the Case rather than disappearing inside a confident answer.
Human authority
Rabbit can prepare analysis and recommendations. A person decides, accepts risk and authorises action.
Not a chatbot. Not a project-management platform. Not an autonomous decision-maker.
Rabbit prepares decisions. Humans make them.First domain
Construction is where Rabbit is tested. It is not what Rabbit is.
The shared Core is intended to work across professional contexts. Construction supplies the first active professional context and the current external review setting. Work on applying the approach to Business is exploratory.
Read about domains