OpsightOpsight / Research in motionThree views of experimental search
Evidence and scientific reasoning for experimental R&D
Put experimental evidence in context.
Opsight helps R&D teams connect experimental results with scientific context and working hypotheses. The aim is to make it easier to examine what the evidence supports, while keeping the reasoning behind proposed experiments open to review.
The research question
Good experimental evidence is rarely in one place.
Past campaign results, scientific papers and researchers' experience can all matter to an experiment. But they often sit apart, making it hard to see what is known, what is assumed and where evidence is still thin.
As the number of variables grows, that context is easy to lose when planning an experimental search.
Opsight's research brings experimental data and hypothesis-led scientific reasoning into the same optimization process, with recommendations people can inspect and question.
Results are spread across campaigns, records and papers.
Scientific context and working hypotheses need to be made explicit.
Not every plausible explanation is supported by the evidence.
Turn research context into an inspectable search.
Probabilistic optimization helps explore a complex space; hypothesis-led reasoning brings context to the process. Researchers can review the candidate experiments and the thinking behind them.
- 01Evidence + hypotheses
Gather the context
Set the objective and constraints. Add relevant past results, scientific context and working hypotheses.
- 02Inspectable candidates
Examine the explanations
Use probabilistic optimization and scientific reasoning to propose candidate experiments. Review the hypotheses and rationale behind each.
- 03Evidence + open questions
Update the research record
Add new results and revisit the working hypotheses. Keep a traceable account of what was tested, what changed and what remains uncertain.
The value is not just a proposed experiment. It is an inspectable record of what informed the decision, what was tested and what the evidence can and cannot support.
Testing where scientific reasoning adds value.
Our published work compares Bayesian optimization, LLM-only reasoning and hybrid approaches on a photocatalysis benchmark, reporting the limitations as well as the results.
BORA hybrido3 + Bayesian optimization
25.3 ± 2.9
Bayesian optimizationTested baseline
11.6 ± 4.4
µmol h⁻¹ · mean best hydrogen evolution rate
Cissé et al. · Digital Discovery, 2026 · Table 3
Values are mean ± standard deviation across 20 runs.
These benchmark results are part of a broader research program spanning computational studies and noisy, closed-loop experimentation on a robotic laboratory workflow.
Have experimental evidence you'd like to understand better?
Whether you're reviewing results from an earlier campaign or planning a study with a specific hypothesis, we'd be glad to hear what you're working on.
Researching how scientific reasoning can shape optimization.
Opsight builds on research into hypothesis-led optimization and scientific reasoning in closed-loop experiments at the University of Liverpool.
Our aim is to complement domain expertise and laboratory workflows, not replace them. Recommendations and their supporting rationales should be inspectable, while the distinction between a hypothesis and a validated explanation stays clear.
A few practical questions.
Do I need an automated laboratory?
Not at all. Tell us how you currently run experiments and collect results. We can discuss what a suitable automated or manual workflow would require, including any integration work.
Can I start with an existing dataset?
Yes. Describe it in the inquiry form. Useful datasets include the settings for each experiment, measured outcomes and relevant constraints. Feasibility depends on quality, coverage and noise, not just the number of rows.
Does the AI explain its recommendations?
Opsight's hypothesis-led approach records proposals and rationales for review. These are useful for scrutiny, but they do not reveal every model internal or establish that a suggested mechanism is correct before validation.
Can I try Opsight?
We're inviting R&D partners to explore scoped pilots and retrospective evaluations. Together, we'll assess fit, define what to evaluate and learn where the approach is useful.
Start with the evidence you have.
Tell us about the results, context or hypothesis behind a study. We can discuss whether a retrospective review or a focused pilot is a useful place to begin.
No account or dataset upload required.
Have a dataset or working hypothesis in mind? We'd be glad to hear about it.