Six workflows to copy into your agent

Ask the agent to prepare next Tuesday's client agenda. It pulls campaign performance, standout replies, and key events from one unified history. You walk into the call ready.

Use the Outreach Magic skill. Prepare next Tuesday's client agenda for [CLIENT NAME / workspace slug].

Use the installed Outreach Magic skill. Use built-in query presets when they fit. If numbers look stale, sync the database first. Do not invent metrics. Show query output or counts from my local database.

For that workspace, use query presets for the past 7 days (engagement, replies, interested) and include:
- Week-over-week sends, replies, reply rate, and bounce rate
- 3–5 standout replies with a short quote each
- Notable lead status changes and anything tagged urgent
- Deliverability flags (bounce spikes, risky domains)
- The 3 talking points I should lead with on the call

Format as a concise briefing I can paste into my agenda.

What's underneath: Engagement, replies, and interested presets scoped to a workspace.

One lead can live in multiple workspaces, each with its own lead status and events. The agent can update and report by workspace without overwriting global lead context.

Use the Outreach Magic skill. Compare lead status between workspace "[WORKSPACE_A]" and workspace "[WORKSPACE_B]" for the past 7 days.

Use the installed Outreach Magic skill. Use built-in query presets when they fit. If numbers look stale, sync the database first. Do not invent metrics. Show query output or counts from my local database.

Use workspace-scoped data (workspace summary, workspace lead events/status, not webhook volume totals from events alone). Show:
- Leads whose status changed in each workspace
- Same lead_id with different status in A vs B
- A table: lead identifier, status in A, status in B, last event date

End with which workspace view I should use for follow-up and why.

What's underneath: Workspace summary plus workspace-scoped status and events across two slugs.

Your agent can enrich, verify, and modify contacts directly in your root database so every downstream workflow uses fresh contact data.

Use the Outreach Magic skill. Before we launch campaign/workspace "[SLUG]", audit and fix contact data in the root database.

Use the installed Outreach Magic skill. Use built-in query presets when they fit. If numbers look stale, sync the database first. Do not invent metrics. Show query output or counts from my local database.

1. List leads missing email, title, company domain, or LinkedIn URL.
2. Check local SQLite before any paid enrich or email-find API calls.
3. Use lead-enrich / email-finder companion skills only for gaps still missing after the DB check.
4. Write verified fields back to root lead records and summarize what changed.

End with: ready or not ready to launch, and how many records still need manual review.

What's underneath: Local database check first, then companion skills only for gaps.

Slice reply rates by country, seniority, industry, company size - whatever the agent has data for. Find out which segments convert, kill the ones that don't, ask follow-up questions the way you'd ask an analyst.

Use the Outreach Magic skill. For campaign "[CAMPAIGN_NAME]" (last 90 days unless I say otherwise), analyze segment performance from my local database.

Use the installed Outreach Magic skill. Use built-in query presets when they fit. If numbers look stale, sync the database first. Do not invent metrics. Show query output or counts from my local database.

Join lead_attributes to replies/events. Break down reply rate by seniority and country. Output:
- A table: segment, sends, replies, reply rate
- Top 2 segments to double down on and bottom 2 to pause or kill, with one-line rationale
- Flag segments with fewer than 30 sends as low sample

What's underneath: Lead attributes joined to replies and events with sample-size warnings.

For any campaign, find the exact messages that earned replies. The agent groups them by theme and surfaces what is working. Feeds directly into the next sequence.

Use the Outreach Magic skill. Across all campaigns in the past 14 days, find copy that earned positive replies.

Use the installed Outreach Magic skill. Use built-in query presets when they fit. If numbers look stale, sync the database first. Do not invent metrics. Show query output or counts from my local database.

Use the replies query preset or equivalent local data. Group winning subject lines and opening hooks by theme. For each theme include: reply count, a ≤25-word example quote, and source campaign/workspace.

Finish with 3 specific ideas to test in my next sequence.

What's underneath: Reply rows grouped into themed subject and opening analysis.

Email and LinkedIn history merge into one timeline with auto-deduplication so your agent always uses matched profiles and complete context.

Use the Outreach Magic skill. Show everything we know about [LEAD NAME] at [COMPANY] from local SQLite.

Use the installed Outreach Magic skill. Use built-in query presets when they fit. If numbers look stale, sync the database first. Do not invent metrics. Show query output or counts from my local database.

Merge email touches, LinkedIn touches, enrichment attributes, tags, and workspace status into one chronological timeline (dedupe duplicate events). Quote the last real reply before drafting.

Then write a follow-up email that continues that thread, not generic outreach fluff.

If the lead is missing from the database, tell me exactly what to import or connect first.

What's underneath: Merged timeline, events, and attributes with thread-aware follow-up draft.