Concepts
Teach HAL how your business runs.
Signals say what happened. Operating context says what it means for you — the constraints, goals, and bottlenecks that turn generic recommendations into yours.
What it is
Operating context is a structured profile per project — and optionally per account — that you configure once and revisit occasionally. Fields typically cover: what you sell, who you sell to, current bottleneck (support capacity? activation? pipeline?), hard constraints (team size, response-time promises), and near-term goals.
The fields that matter most
- The current bottleneck — one sentence. This single field most changes which recommendation rises to focus.
- Constraints — what the team can't do right now. Recommendations respect them instead of suggesting two hires.
- Definitions — what counts as "activation," "at-risk," "qualified" in your business. Health signals and insights align to your definitions, not generic ones.
How it's used
- Focus — weighs pressure against context: during a launch month, support response beats pipeline polish.
- AI answers — suggestions and deflection speak in your terms and respect your policies.
- Work proposals — drafts target the bottleneck, not generic best practice.
Using it well
- Be specific about the bottleneck. "Support is slow" is noise. "First response time doubled because we're all at a conference this week" changes what HAL proposes.
- Keep constraints current. Stale constraints produce confidently wrong priorities — a two-minute monthly review is enough.
- Update when the business shifts — new pricing, a new hire, a new product line. Each is a context change, not just a note in Slack.
Context everywhere
The same principle extends to customers: identification traits, CRM fields, and tracked events are context about them. Operating context is context about you. HAL reasons over both.