Brief templates can run any agent on your workspace, at the points in a brief's life that you choose. You add an agent trigger to the template, then pick the agent and when it runs. Its insights appear in the Agent Insights panel on every brief made from that template.
The checks your team relies on become part of every brief, so nobody has to remember to run them. A spec check can review deliverables as soon as Pencil extracts them. A review can run again when a scope change lands. The person who creates a brief template sets up its triggers. For an overview of briefs and templates, read about Brief Mode.
What runs on every brief template
Every brief template comes with two Pencil triggers:
Brief Health Check. Reviews the brief for production readiness when brief extraction finishes. It flags missing or incomplete reference assets, unclear or contradictory instructions and gaps in tone-of-voice direction. You tailor what it checks with the user prompt.
Scribble. Picks up the brief when it's accepted, produces the requested work and adds it to the brief's deliverables.
You can't remove these two triggers or change their agents. You can switch either one off, and you can change when the Brief Health Check runs. Your existing templates work exactly as before until you add triggers of your own.
Brief extraction only runs when you upload a brief document. If your team builds briefs from scratch, add manual creation to the Brief Health Check's trigger points, or it won't run on those briefs.
How do I add an agent trigger?
Start with an agent on your workspace. It can be a Pencil agent or one your team built. If you build an agent for briefs, give it the Brief category so it sits with the other brief agents. For help building one, read creating custom agents.
Then set up the trigger:
Template. In Brief Mode, select Templates, open the brief template and go to its agent triggers page.
Add trigger. Select Add trigger.
Name. Give the trigger a name. It can differ from the agent's name, so you can describe what it does on this template.
Description. Add a short description. People see it on the brief before the agent runs, so it tells them what's coming.
Agent. Choose any agent on the workspace. You can also choose Scribble and give it the task directly.
Trigger points. Pick one or more moments for the agent to run.
User prompt. Optionally, add instructions that tailor the agent to this template.
Notifications. Optionally, choose who hears about it when the agent runs: the brief creator, the brief submitter or the brief recipients.
Choosing trigger points
You can run an agent at any of these moments:
Brief statuses. Any status from submission through to completion or cancellation.
Extraction. When brief extraction finishes, or when deliverable extraction finishes.
Manual creation. When someone builds a brief from scratch instead of uploading a document.
Scope change. When a scope change is applied to an accepted brief.
Pick the moment the agent's input is ready. A spec check against deliverables belongs on deliverable extraction, because the deliverables need to exist first. For more on how Pencil builds deliverables, read about deliverables.
Using one agent on several triggers
The user prompt sits on top of the agent's own instructions. It lets you tailor one agent to a particular brief type without building a new agent. The same agent can sit on several triggers, each with its own user prompt and trigger points.
Managing triggers
You can switch any trigger off without losing its setup. You can also duplicate a trigger to make a variation, or delete a custom trigger you no longer need.
What agents show on a brief
Each trigger on the template appears as a card in the Agent Insights panel on the brief. Before the agent runs, its card shows what the agent does and when it's due to run.
Once the agent runs, the card shows its response. What that contains depends on the agent's instructions. A spec-checking agent, for example, might summarise the changes each deliverable needs, ask questions where it isn't sure, and flag deliverables the spec sheet doesn't cover.
Agents review the brief and recommend changes. You decide what to change.
If the trigger has notifications set up, the people you chose get an Inbox or email notification when the agent runs, depending on their notification preferences. Anyone with the brief open can see the response on the card at any time.
Running an agent on demand
Every agent card has Run now. Use it to run an agent before its trigger point, to run it again after you change the brief, or to retry a run that failed.
If the agent works on deliverables, wait until they finish generating before you run it early.
Ideas for brief agents
Any check or task that helps at some point in the brief lifecycle can become a trigger. A few to get you started:
Master spec analyser. Checks each deliverable against a client's master specifications as soon as Pencil extracts the deliverables. Upload the spec sheet to the agent's knowledge base, then tell it how to compare and what to report.
Asset re-use finder. Looks for similar assets already on the workspace, so your team can re-use work and deliver faster and more cheaply.
Airtable project creator. An Airtable-connected agent that creates a project in Airtable when a new brief comes in.
Performance agent. Checks deliverables against ad performance data and recommends extra deliverables that have performed well in other markets.
Effort agent. Reviews deliverables and scores them in effort points.
Claims agent. Reviews the brief and its deliverables against your claims guidance.
Common questions
Do agent runs use credits?
Yes. Each run uses credits, priced like other agentic text work. For the detail, read how credits work.
Why hasn't my agent run yet?
Agents run at the trigger points set on the template. The agent's card in the Agent Insights panel shows when it's due. If the agent runs on brief extraction, it only runs on briefs created by uploading a document. Select Run now if you don't want to wait.
What happens if a run fails?
Select Run now on the agent's card to retry it.
Can an agent edit the brief for me?
No. Agents review the brief and its deliverables and recommend changes. You apply any changes yourself.
Can Scribble run as a trigger?
Yes. Choose Scribble as the agent and use the user prompt to tell it what to do. Scribble needs the right skills for the task, as any agent does.