Nine reports sit one click away: as chips under the composer on Home and in an empty full-page chat, and as slash commands in any chat. Each one sends a fixed question to the agent, which reads your account's numbers and writes the report.

The nine reports

To use a slash command, type / in the composer, pick the command and send. It sends the same question as the chip.

Rules every report follows

• 7 days by default. One day is treated as noise. Comparisons use your own history, including the same weekday in past weeks, never industry benchmarks.
• No verdict on anything too new. Anything edited in the last 72 hours is shown as in observation, not judged. A verdict on an ad also needs 7 or more days and 30 or more conversions; triage and the daily pulse skip anything with under 3 days of data.
• Metrics, not images. The agent reads names, formats and numbers. It doesn't look at your images or videos, and it doesn't critique copy.
• One account at a time. If you have more than one ad account connected and you haven't said which, the agent asks first. With a single account it just uses it.
• Nothing is changed. Reports only read. If you want to act on one, ask the agent in the same chat.

Triage the account

Asks "What needs my attention today?". You get at most 7 items, ranked by confidence and size. Each item names the campaign, ad set or ad, gives the evidence across 24 hours, 3 days and 7 days, a likely cause, a recommended action and a confidence level. On a quiet account you get one line saying nothing needs you, plus the single most useful improvement.

Scale the winners

Asks "What should I scale today?". Each candidate gets a label:

• Scale now: efficient, stable across 3 and 7 days, with real volume (7 or more days and 30 or more conversions, or spend of at least twice the target CPA) and no fatigue.
• Scale cautiously: winning, but on thinner evidence.
• Do not scale yet: volatile, tiring, too new, or carried by one large order.

Every recommendation comes with a method (for example, raise the budget 15 to 25% on a mature winner) and a pullback threshold that tells you when to reverse it.

Recap the last 24h

Asks "What changed in the last 24 hours?". Each meaningful change is compared with the day before, the 3 and 7 day averages and the same weekday last week, then labelled: normal volatility, positive signal, negative signal, operationally caused, tracking anomaly, needs more time or actionable now. Tracking anomalies come first. The agent checks your team's activity log for launches, pauses and edits before it blames performance. A quiet day gets one line.

Find creative winners

Asks "Find my creative winners.". Winners are grouped by type, such as volume, efficiency, emerging or durable winners, or winners at risk of fatigue. Each shows the evidence, the winning angle (read from the ad's name, format, call to action and metrics), the next move and the fatigue risk. A confident winner needs 7 or more days and 30 or more conversions; anything thinner is labelled emerging.

Write a client update

Asks "Write a client-ready summary.". You get three parts: a suggested subject line, a summary you can send (performance, what drove it, creative, actions taken, next plan) and internal notes with the caveats you shouldn't send. If you haven't given the agent a target, it either asks you for one or says plainly that the summary is based on your recent averages.

Forecast month-end

Asks "Forecast my month-end pacing.". You get projected month-end spend, conversions, revenue, CPA or CPL and ROAS; current pace against required pace; the gap to budget and goal; campaign-level pacing; and suggested budget shifts. If you haven't shared a monthly budget or goal, the agent asks for it and forecasts on your current run rate meanwhile, labelled as such.

Today's Pulse

Runs the daily briefing. It's meant to be read in under a minute and has up to six parts: a status line, a performance snapshot of your main KPI, budget pacing, up to 3 creative highlights, up to 2 suggestions and changes from the last 72 hours. Parts with nothing to say are left out, so a healthy day gets a short pulse.

Audit creative fatigue

Asks the agent to audit creative fatigue and surface ads that need a refresh. This one isn't a fixed report template: the agent scores your ads with its fatigue check, which rates each one no fatigue, early, active or exhausted, then explains which ads to refresh.

Team report

Asks for a team activity report for the last 7 days. The agent reads your team's activity log and summarises who did what. When it looks at the results of a change, it only judges changes older than 72 hours.