Skip to content

Services / Research & R&D

Research & Experimental Development

Structured investigation where the technical answer is not known in advance: define the uncertainty, test candidate approaches and document what works.

The approach

When nobody can say in advance whether a technical approach will work, the sensible next step is a structured investigation rather than a production build. KRATT defines the uncertainty, tests candidate approaches and records the evidence.

Problems we are brought in for

  • A concept that depends on a technical capability nobody has proven yet.
  • Accuracy or performance requirements that existing methods do not meet.
  • Several possible approaches and no evidence about which will work.
  • A need for documented technical evidence before committing investment.

Capabilities

  • Technical feasibility studies
  • Systematic investigation of technical uncertainty
  • Proof-of-concept development
  • Experimental software
  • Measurement and validation systems
  • Prototype evaluation
  • Technical research reporting

Typical outputs

  • Defined objective and statement of technical uncertainty
  • Experiment plan with evaluation criteria
  • Prototypes and recorded experimental results
  • Technical report with conclusions and recommended next steps

Research is scoped and recorded differently from ordinary implementation. The research page sets out the process, areas of expertise and recordkeeping approach.

Applied Research & Experimental Development

Kratt Ventures Limited is an Approved Research Provider under New Zealand's Research and Development Tax Incentive (RDTI) framework. Our main research area is AI, machine learning and computer vision.

Official list of Approved Research Providers

Customer and project eligibility for RDTI depends on the activities performed, the customer's circumstances and eligible expenditure. Approved Research Provider status does not make every KRATT service or invoice RDTI-eligible.

Have a technology problem that needs ownership?

Tell us what is not working, what you are trying to build or what you need to understand.

Discuss a research problem