# Recipes

> Questions and workflows to run with the Gensiv MCP server, from a weekly visibility briefing to finding the content your site is missing.

Paste these prompts into your assistant as they are, or change them to fit. Each one assumes you've [connected the server](/docs/mcp/setup) and the assistant knows which brand you mean. If you track several brands, name the brand in the prompt.

## Weekly visibility briefing

A summary of the last 7 days you can paste into Slack.

**Tools used:** `visibility_overview`, `trends_movers`, `citation_domains`, `sentiment_overview`

```text
Give me a visibility briefing for the last 7 days:
- Our visibility and share of voice against each competitor, compared with the 7 days before
- The prompts where we gained or lost the most
- Any new domains AI engines started citing
- Whether sentiment moved

Keep it short enough to paste into Slack.
```

The assistant reads the overview for both weeks, pulls the biggest movers and returns a short summary with the numbers behind it.

## Who beats us on each engine

Find the engines where a competitor shows up more than you.

**Tools used:** `brand_engines`, `visibility_heatmap`, `visibility_overview`

```text
Which competitors beat us on each AI engine this month?
Show a table of engines by brands with visibility, then tell me
the one engine where we lose most.
```

The heatmap with brands as rows and engines as columns answers this in one call. The assistant follows up with the overview for the engine you lose on.

## Why a competitor wins a prompt

Read the actual answers behind a visibility number.

**Tools used:** `prompts_list`, `prompt_insights`, `responses_list`, `response_get`

```text
Find the prompt where our visibility is lowest compared with competitors.
Show me how each brand ranks on it, then read the latest answers and
explain why the engines prefer the competitor.
```

The assistant ranks the brands on that prompt, opens the most recent answers in full and points to what they say about each brand and which pages they cite.

## Sources to win

List the sites and threads that shape AI answers in your category.

**Tools used:** `citation_domains`, `citation_urls`, `reddit_threads`

```text
Which third-party sites and Reddit threads do AI engines cite most
for our prompts? Mark the ones that mention competitors but not us,
and suggest where we should get covered first.
```

You get a ranked list of domains and threads with citation counts, split into your own site, competitors and everyone else.

## Find and fill content gaps

Go from "where are we missing?" to a plan for what to write.

**Tools used:** `content_gaps_overview`, `content_gap_prompts`, `content_gap_get`, `fanout_queries`

```text
Find the 3 prompts our site covers worst. For the worst one, show the
searches AI engines ran, what our site is missing and the recommendation,
then outline an article that would close the gap.
```

The assistant reads the coverage scores, opens the worst gap and drafts an outline. Writing the article itself happens in Gensiv for now.

## Check plan usage

See how close you are to your limits before adding prompts or brands.

**Tools used:** `usage_get`

```text
How much of our plan have we used this month, and which limit
will we hit first?
```

## Add prompts for a new topic

Needs a connection with changes allowed.

**Tools used:** `topics_create`, `prompts_create`, `prompts_set_topics`, `prompts_list`

```text
Create a topic called "Pricing questions" and add 5 prompts a buyer
might ask an AI assistant when comparing tools on price. Don't name
our brand in them. Show me the prompts before you add them.
```

The assistant drafts the prompts, waits for your OK, then adds the topic and the prompts and files them under it. New prompts count toward your plan's prompt limit.

---

Source: https://gensiv.com/docs/mcp/recipes