Creating an agent

The agent wizard is a six-step flow that takes you from "I have an idea for an automation" to a deployed, scheduled agent in a few minutes.

Start from a template, or from scratch

Open Agents and click Deploy agent. The split button offers:

• Deploy custom agent — drops you into the wizard with empty defaults.
• Deploy from a template — opens the Templates view. Pick a template and the wizard opens prefilled with that template's name, prompt, schedule, tone, capabilities, and (for org templates) model and run limits. You can edit anything before deploying.

If you launch from an org template marked locked (the admin disabled customization), there is no wizard — clicking the template deploys the agent in one step with the admin's exact configuration. See /docs/agents/templates for the locked-vs-editable distinction.

The wizard has six steps shown in the left rail:

1. Identity — Name and task
2. Schedule — When it runs
3. Tone — How it writes
4. Capabilities — Tools it can use
5. Engine — Model and limits
6. Notifications — Get told when it runs

Use Continue to advance, Back to step backwards, or click any completed step in the rail to jump straight to it. The final step's button reads Deploy agent (or Save changes when you're editing an existing agent).

Step 1 — Identity

The headline is Give your agent an identity. Three fields:

1. Agent name (required, 120 char max) — e.g. Daily morning briefing. The inline pill checks availability as you type: Name is available or You already have an agent with that name. Names are unique per user.
2. Short description (optional, 500 char max) — one line explaining what this agent does.
3. Task prompt (required, 8,000 char max) — the agent's marching orders. Be explicit. The agent can't ask for clarification mid-run, so include inputs, outputs, and constraints in the prompt itself. The example placeholder reads:

Scan unread client emails. Group them by urgency (reply today / this week / FYI) and save the brief to OneDrive Strata/Briefings as morning-brief.md.

A character counter under the textarea shows usage out of 8,000.

While you're on step 1, Strata asks the AI in the background to suggest the minimum set of capabilities for your task. The suggestion lands on step 4 as a banner with Apply recommendations / Not now buttons.

Step 2 — Schedule

The headline is Set the schedule. Pick a frequency, a time, and a timezone. The options in Frequency are:

• Manual only (no schedule) — agent only runs when you click Run now or call its webhook. The schedule sentence reads Your agent has no schedule. It only runs when you click Run now or call its webhook.
• Every N minutes — pick any cadence from 1 to 59 minutes. Your org's minimum-interval rule still applies.
• Every 15 minutes, Every 30 minutes, Every hour, Every 6 hours — fixed cadences.
• Every day — pick the hour and minute.
• Weekdays only — Mon–Fri at the picked time.
• Specific days — pick weekdays plus hour and minute (defaults to Mon–Fri).
• Custom (advanced) — type a raw five-field cron expression.

Underneath the friendly controls, an expandable Use a raw cron expression disclosure lets you override any friendly selection with a raw cron. Below that, the schedule sentence updates live to summarize what you picked (e.g. "Your agent will run weekdays at 09:00 in America/New York with no end date.").

Click Add an expiration date to set a date after which the agent stops running automatically. Type a YYYY-MM-DD value or pick one from the calendar. Backspace clears it.

See /docs/agents/schedules for cron details, minimum-interval enforcement, and how the next-run time is computed.

Step 3 — Tone

The headline is Choose a tone. Pick one of five writing tones — the agent writes its output in that voice:

The sample updates live as you switch tabs.

Step 4 — Capabilities

The headline is Grant capabilities. Toggle on only the tools the agent actually needs. Capabilities are grouped:

• Database — Run SQL queries (the run_sql tool).
• Outlook — Search inbox, Read full emails, Read attachments, Send email (the Send email toggle is flagged Sensitive because agent runs bypass the chat-time confirmation prompt).
• OneDrive — Search files, Read files, Save files (the Save files toggle is flagged Sensitive).
• Jira — Search issues (the search_jira tool: issues, service requests, and service-desk queues) and Create issues and comment (the write_jira tool, flagged Sensitive — an agent run has nobody to show an approval prompt to, and it only works if your organization allows Jira writes).
• Confluence — Search pages (the search_confluence tool) and Create and edit pages (the write_confluence tool, flagged Sensitive, and likewise gated on your organization allowing Confluence writes).
• Generation — Generate Excel, Generate PDF.
• Code & composition — Analysis (sandboxed), Invoke another agent (calls another agent you own as a sub-task; max nesting depth 3).
• Web — Search the web (live lookups for current events, prices, and documentation via Brave Search).

A group whose extension you are not allowed is hidden entirely rather than shown and refused later, so your wizard may show fewer groups than the list above. Jira and Confluence are two separate groups over one Atlassian connection — an admin can allow one and deny the other, and only the allowed one appears.

A running counter at the top reads "N of M tools enabled". If the AI suggested capabilities on step 1, an AI recommendation banner appears here with the list and Apply recommendations / Not now buttons. Apply replaces the current selection with the AI's set; Not now dismisses the banner.

See /docs/agents/permissions for a deeper dive on safe defaults, the dry-run sandbox, and admin-level controls.

Step 5 — Engine

The headline is Pick the engine. Three sections:

• Model picker — choose the model the agent runs on. The list is drawn live from your organization's Model Catalog, so you see exactly the models your admin has enabled — nothing more. Older generations sit under a Legacy models disclosure (still supported, not recommended for new agents). If an agent is currently pinned to a model your org no longer offers, it stays visible so you can switch it. Click the info icon by the heading to open the AI model costs modal, which lists each enabled model with its provider and live per-million-token input/output rates. For the full live catalog and how models are governed, see Models.
• Response mode — Auto, Technical, or Simple. Same modes as chat.
• Response size — three preset tiles:
  • Quick — short answers (60s runtime, 5,000 tokens).
  • Standard — recommended for most agents (300s runtime, 20,000 tokens).
  • Thorough — deep analysis or long reports (600s runtime, 60,000 tokens).
  • Expand Advanced — set the exact numbers to type runtime (30–600 seconds) and max-tokens (500–100,000) directly.
• Active on deploy — toggled on by default. Turn off to deploy in a paused state and resume later from the Agents list.

Note

The picker lists the models your organization has enabled. The model an agent is actually allowed to run on is the intersection of what's enabled for your organization, your role, and any per-user model access your admin set for you — Strata enforces that intersection at run time. An agent always runs on its pinned model; if an admin later disables (or otherwise removes access to) that model, the agent's scheduled runs fail with a clear message until you switch it to an allowed model.

Step 6 — Notifications

The headline is Get told when it runs. The main toggle is Email me when this agent runs (off by default).

Once on, the configuration expands:

• When to email — After every run or Only on failures.
• Send to — defaults to your connected mailbox if Outlook is connected; otherwise type any email.
• CC / BCC — comma-separated email lists.
• Subject template — supports placeholders {agent.name}, {agent.id}, {status}, {elapsed}, {output}, {error}.
• Message template — same placeholders.
• Include the run output in the email — off → just status; on → the agent's output (truncated to 12,000 characters) is included.

See /docs/agents/notifications for placeholder details, the system-sender identity, and deliverability tips.

Click Deploy agent

The final step's button is Deploy agent (or Save changes when editing). On click, Strata validates each step, then POSTs to /api/agents. On success the wizard closes and you're returned to the Agents list with a toast titled Agent deployed reading <name> will run on schedule. (when editing, the toast reads Agent updated / Changes saved.).

What can go wrong

• Name and task prompt are required. — step 1 validation. Fill in both fields.
• You already have an agent with that name. Pick a different name. — names are unique per user (case-insensitive). The inline pill catches this while you type; if the name was taken between typing and clicking Deploy, you'll see the toast version of the same message.
• Pick at least one day of the week. — you chose Specific days but didn't select any.
• A cron expression is required. — you switched to Custom (advanced) but left the cron field blank.
• Invalid cron expression: ... — the cron string didn't parse. Use five fields (minute hour day-of-month month day-of-week), e.g. 0 9 * * 1-5.
• Agents are disabled for your role or organization — your role's canCreateAgents is off, or the org-wide lock_agents flag is on. Contact your admin.
• Agent creation has been disabled for your account by your administrator. — your admin set the per-user lock. You can still pause/resume and star existing agents; you can't create new ones or edit existing prompts/schedules until the lock is removed. See /docs/agents/per-user-governance.
• Agent quota reached: your organization allows up to N agent(s) per user. Delete an existing agent to make room. — you've hit your per-user max. Delete an unused agent and retry.
• Schedule is too frequent: your organization requires at least Ns between runs... — your schedule fires more often than the org's minimum interval. Pick a less frequent cadence.

Related

• /docs/agents — Agents overview.
• /docs/agents/schedules — schedules in detail.
• /docs/agents/notifications — notifications in detail.
• /docs/agents/webhooks — webhook triggers.
• /docs/agents/permissions — capabilities and safety.
• /docs/agents/templates — templates.
• /docs/chat/models — the Model Catalog and how models are governed.