Intelligence · Glossary
Working definitions.
Short, quotable definitions of the operating, AI, and method terms we keep using. Linked to the pillar where each one is explored in depth. Built so search engines and LLMs can ingest the lot in one pass.
AI operating system
also: AI OS, AI based operating system, AI powered operating system- The connective layer between AI models and the work a company does. An AI OS coordinates data, tools, agents, governance, and the operating cadence around them so AI capability turns into compounding output rather than isolated demos. Read more →
Operating model
- A documented description of how a company is organised: structure, roles, decision rights, and key flows. An operating model is intent. An operating system is what actually runs. Read more →
Operating intervention
- The smallest deliberate change to an operating system that produces the largest second-order effect. Craft is in choosing the smallness, not the scope. Read more →
Read · Craft · Scale
- Deepgrain's three-movement method. Read the grain of the existing operating system. Craft the smallest intervention that fits it. Scale the intervention into a repeatable pattern without breaking what works. Read more →
The grain
- The pre-existing direction of how an organisation actually operates: who decides, what compounds, what gets ignored. Cut with it and the work compounds. Cut against it and you spend the rest of the year sanding. Read more →
Five pillars of AI readiness
- Data, tools, agents, governance, and operating cadence. The five surfaces that determine whether a company can absorb AI as capability rather than as demo. Read more →
AI operating ladder
- Five tiers of AI operating maturity: ad-hoc, assisted, augmented, autonomous, and (rare) fully self-operating. Each rung is a different shape of AI operating system underneath. Read more →
Operating cadence
- The rhythm at which an operating system is reviewed, maintained, and adjusted. Without cadence, even a well-designed AI OS rots in place. Read more →
AI workspace
- The structured layer of custom instructions, projects, reference documents, and shared prompts that turns a generic AI tool into a function-specific colleague. The first artefact most People teams should build. Read more →
Champion model
- A staffing pattern for building AI capability inside a function without engineers. Three or four operators given air cover, time, and a build-first remit, supported by a coaching cadence. Read more →
AI agent
- A model plus a goal plus the ability to take steps. Most companies need three or four agents doing the work that previously clogged three or four roles, not a fleet. Read more →
AI governance
- The set of rules, escalation paths, audit trails, and human-in-the-loop checkpoints that decide what an AI system is allowed to do. Governance is not the brake. It is the steering. Read more →
AI pilot
- A time-boxed proof that a model can perform a task in a specific context. Pilots prove capability. Production proves an organisation can absorb the consequences. Read more →
AI infrastructure
- The persistent, owned layer of data pipelines, tool integrations, governance, and runtime that turns one-off AI experiments into compounding capability. The point at which an AI programme becomes an AI operating system. Read more →
AI workflow
- An end-to-end sequence of work where AI handles one or more steps that previously required judgment. The unit of value most People functions should build against. Read more →
AI readiness
- The condition of an organisation's data, tools, agents, governance, and cadence such that it can absorb AI as capability. Readiness is not a model selection problem. It is an operating problem. Read more →
Operating consultancy
- Consultancy that intervenes in how a company actually runs, not how it presents. Diagnostic depth, craft-level intervention, scaled by the cadence the operating system can sustain. Read more →
Operating reality
- The actual present-tense state of how an organisation runs, separate from the strategy that describes its intended future. Most leadership teams confuse the two. Read more →
AI Operating Index (AIOI)
- Deepgrain's free 8-pillar diagnostic that scores an organisation's AI operating maturity across data, tools, agents, governance, cadence, leadership, talent, and intervention readiness. Read more →
Founder mode
- The leadership posture of operating with founder-level authority across the system: re-deciding from first principles, ignoring formal handoffs. A muscle, not an identity. Read more →
Operator mode
- The leadership posture of compounding within a system: protecting interfaces, paying down operating debt, scaling what works. The companion muscle to founder mode. Read more →
Operating debt
- The accumulated cost of decisions deferred at the operating-system layer: stale data ownership, unmaintained workflows, governance gaps. Compounds quietly until the system can't absorb a new initiative. Read more →
AI native
- An organisation designed around the assumption that agents, not just employees, perform work. Different roles, different ratios, different leverage from AI-bolted-on functions. Read more →
People Ops AI Brain
- Deepgrain's working library of nine notes on running a People function with AI: domain map, workflow framework, champion model, governance, value measurement, and more. Read more →