
Agent Engine

1
Brief
Describe the objective — target accounts, signals, quality standards, approval rules. Plain language.

2
Plan
The Engine translates the brief into an execution plan. Every step inspectable before it runs.

3
Execute
Agents read from the Ontology, enrich through the Waterfall, and produce outputs — emails, scores, briefs, CRM updates.

4
Approve
Every output lands in an approval queue with full context. Nothing executes without sign-off unless explicitly auto-approved.

5
Compound
Outcomes feed back into the logic. The agent sharpens over time. Every adjustment is auditable.
Codebase
Open-source (Doghouse). MIT license. Full source on GitHub.
Interface
Natural language briefs translated into inspectable execution plans. No visual workflow builder required.
Templates
90+ pre-built agents covering every revenue motion. Production-ready with Ontology reads, Waterfall calls, and CRM writes. Deploy in a day, customize in an hour.
Execution model
Persistent agents with continuous monitoring, state management, retry logic, and parallelization. Not trigger-and-forget.
Shared context
Every agent reads from and writes to the Revenue Ontology. No integration layer between agents. What one learns, all know.
Approval gates
First-class runtime primitive. In-context approval with full reasoning, data provenance, and inline editing. Approval patterns feed back into agent logic.
Audit trail
Full execution trace per run — every data read, every condition evaluated, every branch taken, every output produced. Readable by ops, not just engineering. Exportable.
Learning
Outcome-driven optimization. Trigger weights, content variants, provider sequences, and scoring models adjust based on measured results. Every adjustment logged with rationale.
Deployment
Single-tenant. Runs in your environment. SOC 2 Type II. AWS PrivateLink.
Data Waterfall
150+ enrichment providers. Sequential routing optimized per segment. The best answer wins. No vendor lock-in.

Agent Engine
Open-source execution engine. Workflows defined in code. Human-in-the-loop checkpoints. Full audit trail on every action.

Revenue Ontology
Every data source normalized into one model. Entity resolution across systems. Relationships stored, not inferred. Schema that evolves with your business.

What is the Agent Engine, exactly?
Why open-source?
How long does it take to deploy an agent?
What happens when an agent makes a mistake?
Can our engineering team build custom agents?





