Yolando

View saved version →

---
name: yolando-brand-intelligence
description: >-
  Use the Yolando MCP connector to analyze a brand's visibility in AI search
  (ChatGPT, Claude, Gemini, Perplexity, and other AI answers), pull competitive
  intelligence, review marketing recommendations, and manage tracked topics,
  prompts, and competitors. Trigger when the user asks how their brand shows up
  in AI answers, about AI-search share of voice, citations, brand reputation in
  LLMs, competitor tracking, or anything in their Yolando workspace. Do not
  trigger for general SEO/keyword questions unrelated to Yolando or AI-answer
  visibility.
---

# Yolando brand intelligence

Yolando tracks how brands show up in AI-generated answers — discoverability,
share of voice, citations, and reputation across AI platforms — plus competitor
intelligence and marketing recommendations. This skill is the playbook for using
the Yolando MCP tools well: which tool to call, in what order, and the pitfalls
to avoid. The tool schemas themselves come from the connector; this tells you how
to drive them.

## Always target a workspace first

A Yolando connection can span several workspaces. Every data tool takes an
optional `workspace_id`.

1. If you don't already know the workspace, call **`list_workspaces`** first. It
   returns the workspaces this connection can act on, with their IDs and names.
2. Pass the chosen `workspace_id` to every follow-up call, and reuse the same one
   across the session.
3. If the connection covers exactly one workspace, you may omit `workspace_id`.
4. If it covers several and the user hasn't said which, ask — or match the name
   they mention (e.g. "the Acme workspace") to a `list_workspaces` result. If a
   call returns `WORKSPACE_REQUIRED`, that means you must supply `workspace_id`.

## Pick the right tool for the intent

**AI-search visibility & insights** — how the brand shows up in AI answers:

- `get_insights_overview` — the top-level visibility summary; start here for
  "how am I doing?"
- `get_platform_insights`, `get_platform_insights_by_brand`,
  `get_platform_insights_by_topic` — break visibility down by AI platform, by
  brand, or by topic.
- `list_prompt_insights` — per-prompt performance (which tracked prompts drive
  mentions).
- `get_citation_insights` — citation share and which domains get cited in AI
  answers.
- `get_reputation_overview` — brand reputation/sentiment by theme.

**Competitive intelligence** — how rivals are doing and what's changing:

- `list_competitors` — the tracked competitor set.
- `get_competitor_discoverability_trends`, `get_competitor_citation_trends`,
  `get_competitor_reputation_trends` — competitor trends over time; use to
  compare against the user's brand.
- `list_radar_feed`, `get_radar_item` — the competitive-intelligence radar
  (market signals and competitor site changes).

**Marketing & content**:

- `list_recommendations`, `get_recommendation` — Yolando's prioritized
  recommendations; rank by priority/confidence when presenting.
- `list_content`, `get_content_item` — content pieces and their status.
- `list_brand_voices` — configured brand voices.
- `get_website_vitals` — site health/technical metrics.

**Workspace structure (reads)**: `list_topics`, `list_prompts`,
`list_prompt_executions` — the topics, prompts, and prompt-run history that the
insights are built on.

## Common workflows

- **"How is my brand doing in AI search?"** → `get_insights_overview`, then
  deepen with `get_platform_insights` (where you're strong/weak),
  `get_citation_insights` (who cites you), and `get_reputation_overview`
  (sentiment). Summarize; don't dump every field.
- **"How do I compare to competitors?"** → `list_competitors`, then the relevant
  `get_competitor_*_trends`, and set them against the user's own trend.
- **"What's new / what changed?"** → `list_radar_feed`, then `get_radar_item`
  for anything notable.
- **"What should I do?"** → `list_recommendations`, ranked by priority and
  confidence; pull detail with `get_recommendation`.

Prefer the narrowest tool that answers the question. Use the time-window and
limit arguments on the insights tools to keep responses focused, and paginate
`list_workspaces` rather than assuming a single page.

## Reads are safe; writes change the workspace

Everything above is read-only. Yolando also exposes write tools that modify the
workspace and require the `mcp:write` scope alongside the relevant domain scope:

- Create/update: `create_topic`, `update_topic`, `create_prompts`,
  `update_prompt`, `batch_update_prompts`, `create_competitor`,
  `update_competitor`, `update_brand_voice`, `create_content_piece`, and the
  skill-registry / chat-suggestion authoring tools.
- Destructive: `delete_topic`, `delete_prompts`, `delete_competitor`,
  `delete_brand_voice`, `delete_skill`, `delete_skill_resource`,
  `delete_chat_suggestion`, and `archive_radar_item`.

Rules for writes:

- Only call a write tool when the user has clearly asked to change something.
- **Confirm before any destructive call** (`delete_*`, `archive_*`) — state
  exactly what will be removed and get a yes.
- `create_competitor` and `create_content_piece` also kick off background
  work (a competitor crawl, content generation); tell the user it's been
  started rather than implying instant completion.

## When something fails

- **`WORKSPACE_REQUIRED`** → you didn't supply `workspace_id`; call
  `list_workspaces` or ask which workspace.
- **Insufficient-scope / permission error** → the connection wasn't granted the
  scope that tool needs (e.g. writes need `mcp:write`). Tell the user to
  reconnect and grant the additional scope; don't retry blindly.
- **A feature/tool is unavailable** → it may not be included in that workspace's
  Yolando plan. Say so instead of guessing an answer.

Referenced files: 1