01 · Define
Agree the user need, product format, intended outcome and a realistic comparator. Decide what evidence would change the development decision.
Our development approach connects research interests to formulation realities. The starting point is a specific problem; the decision to advance depends on evidence.

Agree the user need, product format, intended outcome and a realistic comparator. Decide what evidence would change the development decision.
Assess published work and supplier dossiers. Record where the material, concentration or model differs from the intended product.
Translate the brief into a feasible base. Evaluate texture, compatibility, preservation, packaging and cost-in-use.
Agree a proportionate evaluation plan with appropriate specialists. Keep bench observations, human measurements and consumer perceptions separate.
Advance, reformulate or stop against predefined criteria. A promising story does not override a weak result.
Prepare the approved specification, manufacturing instructions, evidence record and claim wording for the intended market.
| Area | Current position | Project implication |
|---|---|---|
| Formulation and manufacturing | Available through the Anadoris group | Discuss scope, facilities and commercial terms through Made by Anadoris. |
| AI-assisted discovery | Exploratory direction | No proprietary discovery engine, benchmark accuracy or novel molecule output is demonstrated here. |
| Advanced biological assays | Subject to specialist collaboration | Confirm methods and laboratory capability with the proposed study partner. |
| Human efficacy studies | Project-specific evaluation route | Confirm protocol, oversight, recruitment and analysis before commissioning work. |
| Precision fermentation | Ingredient and technology partnership interest | No in-house fermentation capability is claimed. |
The proposed role of computation is to help organise evidence and prioritise questions. It cannot replace material identification, safety assessment or product testing.
A useful candidate record would retain the source, date, tested material, method, observed outcome and uncertainty. Model suggestions should be distinguishable from experimentally observed results.
Read the AI discovery note →