Glossary
Operations and automation terms, in plain language
Short definitions of terms used across OpsGrit, with links to the articles that go deeper.
- AI agent
- Software that uses a language model to decide what step to take next, often by calling tools, to reach a goal. It differs from a fixed workflow, where a person has decided every step in advance. AI agents vs workflows →
- API (application programming interface)
- A defined way for one piece of software to ask another for data or to make it do something, using structured requests and responses. Webhooks, APIs and iPaaS →
- Audit trail
- A record of who did what and when, kept so that actions and approvals can be reviewed later.
- Automation
- Using software to perform a task or a sequence of tasks without manual effort each time it runs.
- Change log
- A short list at the end of a document that records each version, its date, the author and what changed. SOP governance →
- Dead letter queue
- A holding place for messages or tasks that could not be processed after retries, so a person can inspect and replay them instead of losing them.
- Escalation
- Passing a problem to a more senior or specialized person when the first responder cannot or should not resolve it.
- Human-in-the-loop (HITL)
- A design where a person reviews or approves an automated step, usually where errors are costly or the system is uncertain. Human-in-the-loop approvals →
- Idempotency
- A property of an operation: running it twice has the same effect as running it once. It makes retries and duplicate messages safe. Integration patterns →
- Incident
- An unplanned event that disrupts or threatens a business service, or puts data, money or customers at risk. Incident response runbook →
- iPaaS (integration platform as a service)
- A hosted platform for building and running integrations between applications, usually with a visual builder and prebuilt connectors. Zapier vs Make vs n8n →
- Least privilege
- Giving a person or system only the access it needs to do its job, and no more.
- Low-code / no-code
- Tools for building workflows and simple applications with visual interfaces instead of writing most of the code by hand.
- OCR (optical character recognition)
- Technology that reads text from images or scanned documents, often a first step in invoice and form processing. AI invoice processing →
- Process owner
- The person accountable for how a process performs end to end and for keeping its documentation correct.
- RPA (robotic process automation)
- Software that imitates user actions on screens, such as clicking and typing, to automate tasks in systems that lack good APIs.
- Runbook
- A step-by-step operational guide for a specific system or situation, such as responding to an outage. Incident response runbook →
- Service level agreement (SLA)
- A commitment about how fast a team will respond to or resolve a request, often with consequences if missed.
- Severity level
- A rating of how serious an incident is, used to decide how fast and how widely to respond.
- SOP (standard operating procedure)
- A documented, repeatable way to perform a task so the result is consistent no matter who does it. How to write an SOP →
- System of record
- The system whose version of a piece of data is treated as correct when systems disagree. System of record →
- Tech stack
- The set of software tools and platforms a team or company uses to do its work. Tech stack audit →
- Ticket triage
- Sorting incoming support requests by category, urgency and owner so they reach the right person quickly. AI ticket triage →
- Total cost of ownership (TCO)
- The full cost of a tool or system over its life: licenses, implementation, integration, training, administration and exit. Build vs buy →
- Webhook
- A message a system sends to another system when something happens, instead of waiting to be asked. Webhooks, APIs and iPaaS →
- Workflow
- A defined sequence of steps, with triggers and decisions, that moves work from start to finish.