Use cases

These are recurring problem classes the optimization engine is built for—not a claim of fully packaged vertical suites. Domain-specific libraries remain a limited, evolving surface; the core product is formulate → solve → explain.

Resource allocation

Budget, staff, inventory, and compute across competing objectives with hard caps and minimums. The engine searches feasible regions instead of hand-tuning heuristics.

Portfolio and tradeoff optimization

Balance return, risk, and policy limits when objectives conflict. Useful when “best” is a constrained frontier, not a single closed form.

Planning and scheduling

Routing, timetables, and capacity planning where discrete choices and continuous parameters mix. Sematryx targets feasible, high-quality schedules under stated rules.

Constraint-heavy business decisions

Regulatory, safety, and operational rules encoded as constraints—not afterthoughts. Explanations help teams justify choices to stakeholders and auditors.

Agent-assisted optimization workflows

Agents call the same engine via MCP or API: formulate from natural language, solve with hosted compute, return numbers plus narrative the agent can use downstream.

See benchmark evidence and the full product workflow on the pages below.