> ## Documentation Index
> Fetch the complete documentation index at: https://docs.botdog.co/llms.txt
> Use this file to discover all available pages before exploring further.

# Example Prompts

> Real prompts to try once Botdog is connected — grouped by what you want to get done.

Once Botdog is connected (see the [Quickstart](/mcp/quickstart)), you talk to it in plain language. You never name a tool — the assistant reads your intent and calls the right [tool](/mcp/tools) for you. The prompts below are starting points; rephrase them however you like.

> The tool name shown after each prompt is what the assistant calls under the hood. You don't type it.

## Explore your campaigns

> "Show me all my Botdog campaigns."

→ `list_campaigns`

> "Which of my campaigns are still active?"

→ `list_campaigns` (status filter)

> "Give me the full details and current status of my 'Q2 Founders' campaign."

→ `list_campaigns` → `get_campaign`

> "How is 'Q2 Founders' doing — invites sent, acceptance rate, replies, profile visits?"

→ `get_campaign_statistics`

## Analyze performance

> "What were my outreach numbers over the last 30 days?"

→ `get_analytics`

> "Chart my daily invites and replies from May 1 to May 28."

→ `get_analytics_daily`

> "Break down the last 7 days by team member — who's getting the most replies?"

→ `get_analytics_by_user`

> "Compare my acceptance rate over the last 7 days against the last 30 days and tell me if it's trending up."

→ `get_analytics` (two periods, then the assistant compares)

## Look up your team

> "List everyone on my Botdog team."

→ `list_users`

> "Which of my team members have a LinkedIn account connected?"

→ `list_users` → `get_user` (per member)

> "Show my team's profile and plan."

→ `get_team`

## Add leads to a campaign

Write actions require the `mcp.write` scope (see [Scopes](/mcp/scopes)). They touch a **live** outreach campaign, so the assistant confirms the campaign and the lead list with you before sending.

> "Add these LinkedIn profiles as leads to my 'Q2 Founders' campaign: linkedin.com/in/jane-doe, linkedin.com/in/john-roe."

→ `add_leads_to_campaign`

> "Find my 'Q2 Founders' campaign, then add [john@acme.com](mailto:john@acme.com) and linkedin.com/in/jane-doe as leads."

→ `list_campaigns` → `get_campaign` → `add_leads_to_campaign`

## Work your inbox

> "Who replied to me this week?"

→ `list_replies`

> "Show me only my unread replies."

→ `list_replies` (unread filter)

> "Show unread replies on my 'Q2 Founders' campaign and summarize what each person wants."

→ `list_replies` (campaign + unread filter)

> "Pull up the full record for linkedin.com/in/jane-doe — has she replied, what was her last message?"

→ `get_lead`

## Inspect a campaign's leads

> "List the leads in my 'Q2 Founders' campaign."

→ `get_leads_from_campaign`

> "Which leads in 'Q2 Founders' have replied so far?"

→ `get_leads_from_campaign` (reply filter)

> "Show me the leads that errored out in 'Q2 Founders'."

→ `get_leads_from_campaign` (status filter)

## Build and manage lists

> "Create a new list called 'Hot Replies'."

→ `create_list`

> "Add these profiles to my 'Hot Replies' list: linkedin.com/in/jane-doe, linkedin.com/in/john-roe."

→ `add_leads_to_list`

> "What lists do I have, and how many leads are in each?"

→ `list_lists`

> "Show me everyone in my 'Hot Replies' list."

→ `get_leads_from_list`

> "Remove linkedin.com/in/john-roe from my 'Hot Replies' list."

→ `delete_leads_from_list`

## Control live campaigns

Write actions require the `mcp.write` scope (see [Scopes](/mcp/scopes)) and touch **live** outreach, so the assistant confirms with you first.

> "Pause my 'Q2 Founders' campaign."

→ `pause_campaign`

> "Resume 'Q2 Founders'."

→ `resume_campaign`

> "Stop reaching out to linkedin.com/in/john-roe in my 'Q2 Founders' campaign — he asked to be left alone."

→ `stop_lead_in_campaign`

## Check your sender accounts

> "Which LinkedIn accounts are connected, and are any of them disconnected?"

→ `list_linkedin_accounts`

## Connect your stack with webhooks

> "Create a webhook to POST new replies to [https://hooks.myapp.com/botdog](https://hooks.myapp.com/botdog)."

→ `create_webhook`

> "List my webhooks."

→ `list_webhooks`

> "Point my 'CRM sync' webhook at a new URL."

→ `update_webhook`

> "Delete the webhook I set up last week."

→ `list_webhooks` → `delete_webhook`

## Set up the API in your own code

The MCP server can also explain how to call the Botdog **REST API** from your codebase — base URL, API-key auth, a ready-to-paste snippet, and the list of supported operations.

> "How do I set up the Botdog API in my Node project?"

→ `get_api_setup_guide` (language: node)

> "Give me a curl example for authenticating to the Botdog API."

→ `get_api_setup_guide` (language: curl)

> "Show me a Python snippet to list my campaigns via the API."

→ `get_api_setup_guide` (language: python)

> "What operations does the Botdog API support?"

→ `get_api_setup_guide`

## Multi-step workflows

The assistant can chain tools to answer one question. A few that work well:

> "Find my worst-performing active campaign this month and explain why."

→ `list_campaigns` → `get_campaign_statistics` → `get_analytics`

> "Pull reply rate by teammate for the last 30 days and summarize who might need coaching."

→ `get_analytics_by_user`

> "Audit my account: list active campaigns, show each one's acceptance and reply rate, and flag any below 20%."

→ `list_campaigns` → `get_campaign_statistics` (per campaign)

> "Find everyone who replied to 'Q2 Founders' this week and put them in a new 'Hot Replies' list."

→ `list_replies` → `get_lead` (per reply, for the LinkedIn URL) → `create_list` → `add_leads_to_list`

> "Pause every active campaign with an acceptance rate under 15%."

→ `list_campaigns` → `get_campaign_statistics` (per campaign) → `pause_campaign`

## Ideas to try

Higher-leverage ways teams use the Botdog MCP server:

* **Daily inbox triage.** "Summarize my unread replies and group them by intent (interested / not now / wrong person)." The assistant reads `list_replies` and does the sorting — your morning queue in one prompt.
* **Reply-to-list pipelines.** Turn positive replies into a segmented audience (`list_replies` → `get_lead` → `create_list` → `add_leads_to_list`) you can export, enroll, or hand to sales.
* **Performance autopilot.** "Audit my campaigns and pause anything underperforming" combines analytics with `pause_campaign` so you stop wasting sender capacity on dead campaigns.
* **CRM / Slack sync.** Wire `create_webhook` to your own endpoint and push reply events into HubSpot, a Slack channel, or an internal dashboard — no polling.
* **Compliance & opt-outs.** "Stop outreach to this person everywhere" via `find_leads_by_linkedin_url` → `stop_lead_in_campaign` (once per campaign) keeps unsubscribe requests honored without logging into the app.
* **Sender health checks.** "Are any of my LinkedIn senders disconnected?" (`list_linkedin_accounts`) catches stalled accounts before they tank a campaign.
* **Works from any MCP client.** The same prompts run in Claude, ChatGPT, Cursor, Clay, n8n, or Make — so you can embed these flows into automations, not just chat.

## What's next

* Full input schemas for every tool: [Tools](/mcp/tools)
* What each scope grants: [Scopes](/mcp/scopes)
* Connection issues: [Troubleshooting](/mcp/troubleshooting)