The task language
A compact way to declare roles, facts, rules, state, and permitted actions across domains.
AUGMENTED INTELLIGENCE
Enterprises need agents that hold fluent conversations and carry out work under the organization’s rules. We built Apollo-1 around both requirements.
BUILT FROM REAL WORK
What do they need to know? Which rules apply? When do they ask for confirmation? That research produced the symbolic task language and runtime at the core of Apollo-1.
developing the task language
and runtime
of task-oriented
dialogues studied
human agents behind
the conversations
Research history as described in AUI’s September 2026 materials.
COMPANY FACTS
WHY NOW
A compact way to declare roles, facts, rules, state, and permitted actions across domains.
LLMs interpret open-ended language without requiring the symbolic system to encode the whole world.
Coding agents can write, run, inspect, and revise programs against explicit tests.
ONE RUNTIME / MANY KINDS OF WORK
Production experience spans automotive retail, travel insurance, enterprise software, consumer AI, and ecommerce software.
What neuro-symbolic AI meansStart with a well-defined task. Inspect how Apollo-1 performs against your rules and integrates with your operation.
Explore the technical thesis, evidence, and economics behind a runtime designed to serve many domains through one task language.
PUT APOLLO-1 TO WORK