# Use Cases

You describe the outcome; the assistant figures out which tools to call. Each workflow below chains a few of them together, so a single message does the work of ten dashboard clicks.

These are patterns, not scripts. Swap in your own brands, topics, dates, and platforms. The "Ask" line is what you type; everything under it is what the assistant does and where to take it next.

## Weekly visibility check

**Goal:** a two-minute Monday read on where you stand.

**Ask**  
Summarize our AI visibility for the last 7 days versus the 7 before. Break it down by platform, name the biggest movers, and flag anything that dropped.

**Copy**  
**What it does:** pulls your headline KPIs, the visibility timeseries, and per-platform breakdowns, then writes the recap.

**Take it further:** "Now just ChatGPT and Perplexity." · "Turn that into three bullets I can paste into Slack." · "What changed in our worst topic?"

## Competitor deep dive

**Goal:** see exactly where a rival is pulling ahead.

**Ask**  
Compare our visibility and share of voice against every tracked competitor over the last 30 days. Who wins each topic, and where are we slipping?

**Copy**  
**What it does:** lists your tracked brands, runs the comparison matrix, and breaks it down by topic.

**Take it further:** "Which prompts is [competitor] in that we're not?" · "Show their trend, are they rising or did they just spike?" · "What sources cite them in our top topic?"

## Find your citation gaps

**Goal:** a target list for digital PR and content.

**Ask**  
Show the top domains and URLs cited in answers about our topics over the last 30 days. Which ones mention competitors but never us?

**Copy**  
**What it does:** reads citation analytics by source and domain, then filters for the pages you're missing from.

**Take it further:** "Rank those by how often they're cited." · "Which of these could we realistically earn a mention on?" · "Draft an outreach angle for the top three."

## Sentiment investigation

**Goal:** get past the score to the reason behind it.

**Ask**  
What's the sentiment toward our brand in AI answers right now, and which themes are driving the negative ones?

**Copy**  
**What it does:** reads the sentiment summary and theme breakdown, then groups the drivers for you.

**Take it further:** "Pull the actual negative mentions so I can read them." · "Is this one engine or all of them?" · "Has it shifted since last month?"

## Organic and AI side by side

**Goal:** connect AI visibility to traffic that already exists.

**Ask**  
Compare our Search Console clicks with our AI visibility for the same topics this month. Where do we rank well organically but stay invisible in AI?

**Copy**  
**What it does:** pulls Search Console numbers and AI visibility for matching topics and lines them up.

**Take it further:** "Add AI referral traffic to that." · "Which tracked keywords moved most this week?" · "List pages with the lowest audit scores so I know what to fix first."

## Close the loop

**Goal:** act on a finding without leaving the chat.

**Ask**  
Add a prompt under our 'Pricing' topic for 'best [category] tool for small teams', and start tracking Acme as a competitor.

**Copy**  
**What it does:** calls write tools, so your client asks you to confirm before anything changes. You can also queue agent work: "Create a task to draft a comparison page for that gap, run it, and tell me when it's done."

**Take it further:** "What's the top recommendation Orchly has for us, and turn it into a task." · "Run it and poll until it finishes." · "Show me the document it produced."

## Build your own

Any of these read tools can anchor a workflow. Mix and match:

| You want to know… | Try asking |
| --- | --- |
| Hidden sub-queries | "What does our main prompt fan out into on ChatGPT and Perplexity?" |
| AI referral traffic | "How much traffic are AI assistants sending us this month, and from which?" |
| Rank movement | "Which tracked keywords gained or lost the most this week?" |
| Technical health | "List our tracked pages sorted by audit score, worst first." |
| What to do next | "What are Orchly's top recommendations for us right now?" |
| Raw mentions | "Show recent mentions of our brand in Perplexity answers." |

## Getting better answers

- **Name the workspace** if you run more than one: "in our Acme workspace…". Or just ask it to list workspaces first.
- **Pin the window.** "Last 30 days on Perplexity" beats "lately" every time.
- **Stack follow-ups.** The assistant keeps context, so "now break that down by topic" picks up where it left off.
- **Ask for the format you'll use.** A table, three bullets, a Slack-ready summary, say so up front.
- **Explore without fear.** Reads never change anything; write tools always stop and confirm.

See the full [tools reference](/content/help/mcp/tools/index.html) or grab ready-made [prompts](/content/help/mcp/prompts/index.html).