What counts as an AI agent, and what belongs in the inventory?
An AI agent pursues a goal by choosing its own next step and acting through tools. What counts as one, and what belongs in the inventory.
- Topic
- Agent governance
- Product
- Novie Control
- Format
- Answer
- Reading time
- 4 min
- Published
- 2 October 2026
- Updated
- 21 October 2026
An AI agent is software that pursues a goal by deciding its own next step and acting through tools, such as reading a system, sending a message or changing a record. A chatbot that answers and stops is an assistant. A fixed workflow that calls a model at one step is automation. For an inventory, count anything that can take an action on the organisation's behalf without a person approving that specific action.
A working definition.
Three properties together make something an agent. It is given a goal rather than a single instruction. It decides its own sequence of steps toward that goal, using a model to choose what to do next. And it can act: call tools, use other systems, change things outside the conversation.
Official definitions say much the same. The OECD, in a paper on agentic AI in organisations published in September 2026, describes AI agents as systems that perceive and act on their environment with a degree of autonomy, using tools as needed to achieve specific goals. Singapore's IMDA notes there is no consensus definition, but that agents usually plan, decide and act over multiple steps toward a goal a user sets.
TakeawayAn agent is given a goal, chooses its own steps with a model and acts through tools, and all three together make the difference.
Agent, assistant, automation.
- An assistant answers a person and stops. The person decides what to do with the answer. A chat tool used for drafting is an assistant.
- Automation follows a fixed path someone designed. A workflow that sends each new invoice to a model for a summary and files the result is automation, even though it uses AI at one step.
- An agent chooses its path. Given 'resolve this support ticket', it decides to look up the customer, read the order history, issue a refund and write the reply, and does each of those through tools.
- The line moves. An assistant given tools and permission to use them becomes an agent. A workflow with a model choosing which branch to take is most of the way there.
TakeawayAn assistant answers and stops, automation follows a fixed path and an agent chooses its path, so the line moves as soon as tools and permission are added.
Why the definition matters.
The word is used loosely. Products are relabelled as agents, and some things that act like agents are never called one. An inventory built on the label misses the second group, which is the one that carries the risk.
Ownership is thinner still. In Gravitee's survey of 750 technology leaders, published in June 2026, 7.2 percent of organisations had a named individual formally accountable for AI agent behaviour.
The numbers are moving fast. In KPMG's Q2 2026 AI Pulse Survey, 53 percent of large US enterprises had AI agents deployed and 18 percent were orchestrating several agents across workflows, double the previous quarter. Deloitte's State of AI in the Enterprise 2026 found 21 percent have a mature governance model for agents.
Frameworks written for agents start from the inventory too. Our guide to Singapore's agentic AI governance framework shows what one asks of anyone deploying agents.
TakeawayAgents are deployed in 53 percent of large US enterprises, and 7.2 percent of organisations have a named person accountable for what they do.
What belongs in the inventory.
Use a functional test rather than the vendor's label. Include anything that meets any of these:
- It can take an action in a system (write, send, change, pay, delete) without a person approving that specific action.
- It runs on a schedule or a trigger, without a person starting each run.
- It holds credentials or an API key of its own, or acts under a service account.
- It can call other agents, or be called by them.
- It spends money on the organisation's behalf, through model calls or anything else.
Assistants that only answer a person still belong on the register of AI in use, because of the data they see. They need a lighter entry than an agent does.
TakeawayCount anything that can act, run, hold keys, call other agents or spend without a person approving that specific step.
Edge cases.
- AI features inside licensed software. Include them if the feature can act, such as sending email or updating records, and record the vendor as well as your internal owner.
- Coding agents. Include them: they change code and run commands, often with the developer's credentials.
- Agents in testing. Include them once they hold real credentials or touch real data. A test agent with production access is a production agent.
- Personal agents. An agent a member of staff runs on their own account against company systems is in scope, and is the hardest to find.
TakeawayFeatures inside licensed software, coding agents, test agents with real access and personal agents all belong in the inventory.
How Novie Control does it.
Novie ControlAvailable now
Governance and accountability.
Give every workload a named owner, a budget and clear boundaries.
1,284
calls governed today.
Example figure.
Sources
- OECD, Agentic AI in organisations: Early insights from practitioner interviews, OECD Artificial Intelligence Papers No. 65 (16 September 2026)
- IMDA, Model AI Governance Framework for Agentic AI, Version 1.5 (20 May 2026, updated 5 June 2026)
- Gravitee, State of AI Agent Security Report 2026 (15 June 2026)
- KPMG, AI Quarterly Pulse Survey Q2 2026 (24 June 2026)
- Deloitte, Business and IT leaders report AI agents are scaling faster than their guardrails (24 April 2026)