Operational intelligence, as the Guild uses the term, is the operating layer between a business's data and its results: the decisions that matter, the people who own them, the signals that should trigger them and the actions that follow. Most organisations already have plenty of data and more dashboards than anyone reads. Dashboards are passive. What is missing is the loop that turns information into coordinated action, and the action back into evidence.
The loop has five parts — signals, interpretation, decisions, actions and measurement — and it works at any size. A team of ten can run it on one page and a thirty-minute weekly review. A larger organisation runs the same loop across a portfolio of people and agents.
Questions this topic answers
- What is the difference between a dashboard and operational intelligence?
- Which recurring decisions should we start with, and who owns each one?
- When does a change in a number deserve someone's attention?
- What should happen in the hour after a signal fires?
- How do we know an action changed the outcome, rather than simply happened?
- Where can agents help in the loop, and where must a person decide?
Start here
- Operational intelligence is the missing layer — why dashboards alone do not change outcomes, and the five-stage loop.
- Operational Intelligence for SMBs — Field Guide 033: the loop sized for a small team, with a decision register and a weekly review built on the tools you already have.
- Signal in the noise — Oracle's tests for what counts as a signal, and the shape of a message an owner can act on in a minute.
- The cost of an unowned workflow — finding the queues nobody owns, and putting a number on what they cost.
- Why a good intake system changes the quality of every downstream workflow — every decision inherits what intake got right or wrong.
- The Executive Guide to Agentic Operations — how the loop grows into a portfolio of governed agents with clear ownership.
Key ideas
- Signal — a measure tied to a decision. A number that triggers nothing is context, not a signal.
- Condition — the agreed point at which a signal means "act now", judged against what normal looks like rather than a number picked once.
- Decision owner — one named person accountable for acting when the condition is met.
- Playbook — what happens next: who is told, what evidence they see, the default action and the deadline.
- Decision register — the one-page list of recurring decisions with their owners, signals and playbooks.
- Closing the loop — measuring whether the action worked, and feeding the result back into the signals.
- Outcome over activity — alerts acted on and results changed, not reports produced.
- The operating layer — the owners, approvals and records around people and agents. In the Guild, that is the Citadel.
Who knows this topic
- Oracle — the Guild's analytics and operational intelligence agent. Watches operational data, surfaces what changed, why it matters and who should know, and turns activity into numbers a leadership team can act on. Works inside the Citadel, supervised by a Human Wizard in the Data Wizard role.
- Athena — coordinates the mission, structures discovery and tracks decisions, so a signal reaches the person who owns it.
- Atlas — maps data, integrations and dependencies, so signals come from sources people trust.
- Ares — execution: runs workflows and triggers automation, so that a decision turns into action.
On managed operations the division of labour is short: Sentinel watches, Oracle interprets, humans decide.
Where this knowledge is put to work
If the weekly meeting starts with an argument about which number is right, the problem sits underneath the dashboard. Data & Intelligence builds the trusted data and definitions the loop depends on. Managed operations runs signals, escalations and reviews day to day, with every queue, approval, exception and result visible in the Citadel. If you are not sure which decisions matter most, the Guild Assessment is the place to start.
Have a decision that keeps arriving late? Start the loop there.
