Long-context understanding
Operational work often begins with lengthy or messy source material rather than a clean query.
TECHNOLOGY
Claude as a reasoning layer for operational work.
Lertsma’s architecture is being designed around context, structured outputs, explicit review points, and a clear separation between model reasoning and business workflow logic.
SYSTEM DIRECTION
From raw context to reviewable actions.
This describes the early MVP direction and will evolve as real customer workflows are validated.
1
Input layerDocuments · SOPs · Internal notes · Reports · Customer context
↓
2
Context & orchestrationWorkflow intent · System instructions · Source context · Output schema
↓
3
Claude reasoning layerDocument understanding · Extraction · Multilingual analysis · Structured generation
↓
4
Validation & reviewMissing information · Uncertainty · Human review requirements
↓
5
Operational outputSummary · Key points · Risks · Actions · Follow-up
WHY CLAUDE
Capabilities aligned with the product problem.
Structured reasoning
Lertsma needs defined business outputs such as facts, risks, missing context, and actions.
Multilingual workflows
The initial market frequently mixes Indonesian and English across everyday business operations.
DISCLOSURE
Independent company. No implied endorsement.
Lertsma Labs is an independent startup. References to Claude and Anthropic describe technologies we are evaluating and building with. They do not imply partnership, certification, investment, acceptance into the Claude Startups program, or endorsement by Anthropic.