GetWhys

Query GetWhys and receive a synthesized, evidence-grounded answer drawn from GetWhys's proprietary corpus of buyer interviews and any documents the user's organization has uploaded into GetWhys. The answer is accompanied by credibility signals describing the underlying citations. When the answer cites sources, the query and answer are saved to the user's organization's GetWhys workspace so they can be revisited via the view_in_getwhys link. **One focused question per call.** Each call must answer a single, specific question — do NOT pack multiple sub-questions, comparisons, or "and what about X" clauses into a single `query`. If the task spans multiple sub-topics, issue several parallel calls. Each call returns its own answer; you synthesize across them. Call this whenever the task would benefit from being validated by real buyer evidence or from the user's own org's uploaded research, including: persona pain points / KPIs / decision criteria for a target role × industry × company size; what real customers say about a specific vendor or product; win/loss reasons grounded in buyer voice; real-world pricing, discounting, or packaging mentions; buying-committee composition for a category; the language an actual buyer in a given segment uses; questions about the user's own internal documents that they reference; **validating a draft** — do its claims, pains, and language match what buyers actually say (the validation path when no persona fits and `score_content` is unavailable). Do NOT call for general definitions, technical product documentation, end-to-end content drafting, or questions where no buyer / persona / org-document context is needed. When presenting this answer to the user, include the trailing "GetWhys Sources" paragraph verbatim — it summarizes how the evidence was assembled and is how the user gauges the answer's credibility. Also include the view_in_getwhys link so the user can explore the sources. Do not paraphrase, abridge, or omit them.

query_market_research

Returns the full content of every organization-level (non-project) messaging framework configured in the user's org. Project-specific frameworks are not included. **This is the default messaging-framework tool for content generation.** **Call this when** drafting / writing / producing any content the user expects to publish or send — blog posts, emails, LinkedIn / social posts, ad copy, landing pages, sales pitches, customer stories, internal announcements, scripts — or when adapting / rewriting existing copy. Call it on every content-generation kickoff. Pair with `get_brand_voice` for the same content gen tasks. **Do NOT call** when the user is interrogating or selecting one specific framework — use `list_messaging_frameworks` + `get_messaging_framework` for that. **Do NOT call** when answering a research / market question, summarizing data, comparing competitors, exploring a persona, or doing Q&A, drafting PRDs / product specs, competitive teardowns / battlecards, persona documents, or other internal analytical artifacts. These are messaging *guardrails* on outward-facing publishable copy. These are the user's own configured artifacts.

Return the full attributes of every non-archived organization-level (non-project) buyer persona configured in the user's org, in one response. Project-specific personas are not included. Each persona includes name, description, challenges, motivations, KPIs, key stakeholders, watering holes, interests, business needs, and merged job titles, plus a `persona:<handle>` semantic ID for downstream tool calls. Call this **only** when the user requires the full set. For routine workflows that target one persona, use `list_personas` followed by `get_persona`. These are the user's own configured artifacts.

Returns the user's org-configured brand voice — a set of characteristics, each with do / don't rules that should be applied to written content. **Call this when** drafting / writing / producing any content the user expects to publish or send — blog posts, emails, LinkedIn / social posts, ad copy, landing pages, sales pitches, customer stories, internal announcements, scripts — or when adapting / rewriting existing copy. Call it on every content-generation kickoffs. Pair with `get_all_messaging_frameworks` for content generation tasks. **Do NOT call** when answering a research / market question, summarizing data, comparing competitors, exploring a persona, or doing Q&A about the corpus, when drafting PRDs / product specs, competitive teardowns / battlecards, persona documents, or other internal analytical artifacts — these are prose, but they are not prose-to-publish. These are voice *guardrails* on outward-facing publishable copy. This is the user's own configured artifact.

Fetch the full latest-version content of one organization-level (non-project) messaging framework, identified by `id` (UUID) or `title` (case-insensitive exact match). Project-specific frameworks are not available through this tool. Provide exactly one. **Call this when** the user has named one specific framework. If the user named it verbatim, pass `title` directly — one round trip. If the user's reference is partial / fuzzy, call `list_messaging_frameworks` first and pass `id` afterward. If a `title` call returns a collision error, re-call with one of the returned ids. **Do NOT call this for broad content generation.** When the user wants "on-brand copy" without naming a specific framework, call `get_all_messaging_frameworks` instead — fetching one framework under-applies the org's voice. **Do NOT call** when answering a research / market question, summarizing data, comparing competitors, exploring a persona, or doing Q&A, or drafting PRDs / product specs, competitive teardowns / battlecards, persona documents, or other internal analytical artifacts. This is the user's own configured artifact.

Fetch the full attributes of one organization-level (non-project) buyer persona, identified by its `persona:<handle>` semantic ID. Project-specific personas are not available through this tool. **Call this after `list_personas`** to load the details for the persona you'll target. You can also call it directly when the user names a persona handle verbatim. **Do NOT** guess handles — the call will fail on unknown IDs. Use `list_personas` first to enumerate the available handles. If none fits (empty list, or no match for the audience you need), skip this tool and proceed without persona targeting — don't guess, don't stall. This persona is the user's own configured artifact.

Returns the index of the user's organization-level (non-project) messaging frameworks — each entry is `{ id, title, description }`. Project-specific frameworks are not included. The `description` is the framework's "how to use" note: a short statement of *when* it should be applied — what it's about, plus any specific channel / content type, product, or audience it targets. It may be null. Use these descriptions to judge which framework the user means and to disambiguate similarly-named frameworks before fetching one in full. **Call this when** the user references a single framework with a fuzzy / partial name and you need to disambiguate before fetching it (call `get_messaging_framework({ id })` afterward), or when the user explicitly asks what frameworks are available. **Do NOT call this for content generation.** For drafting / writing prose-to-publish, call `get_all_messaging_frameworks` instead. **Do NOT call** when the user has named a framework verbatim — pass that string directly as `get_messaging_framework({ title })`. **Do NOT call** when answering a research / market question, summarizing data, comparing competitors, exploring a persona, or doing Q&A. **Also do NOT call** when drafting PRDs / product specs, competitive teardowns / battlecards, persona documents, or other internal analytical artifacts. These are the user's own configured artifacts.

Return the lightweight index of the user's organization-level (non-project) buyer personas — just `handle` and `name` per persona, nothing else. Project-specific personas are not included. Constant-size discovery call. **Call this first** whenever a workflow needs to identify a target persona before doing persona-targeted work (content generation, scoring, persona-scoped research). Pick the handle that best matches the target persona, then call `get_persona` for full attributes. **Do NOT call** when the user explicitly asks to see every persona in detail — use `get_all_personas`. **Do NOT call** for content with no target persona. **An empty result is an answer, not an error** — it removes `get_persona` / `score_content`, not the task. Same when nothing in the list matches the audience at hand. Either way, never invent a handle and never stall for setup: continue without persona targeting. The empty response spells out how.

Submit a piece of marketing content and a target persona handle to receive a persona-informed score. Returns an overall score (0–100), dimensional breakdown (what's working, what isn't, why), and specific recommendations. **Call this after drafting** any persona-targeted content — ad copy, email, web copy, blog post — to evaluate how well it will resonate before publishing or sending. Supports iterative improvement: revise based on recommendations and resubmit until the score meets your threshold. **Required revision loop:** Capture the `content_id` after every successful call. For every later revision of the same artifact and persona, pass it unchanged as `content_id`, even after an extensive rewrite or on a later turn. Never reuse it for another artifact or persona, and never invent one. Omitting it starts a new revision chain; separate clients must persist and re-supply it themselves. Each call saves the submitted content and its score to the user's GetWhys workspace as a revision linked by content_id. The org's brand voice is automatically included in scoring when configured. Organization-level (non-project) messaging frameworks are automatically evaluated — the tool selects which frameworks apply based on each framework's description and returns a single pass/fail in `framework_match`. Pass `messaging_framework_id` to also get isolated feedback on a specific organization-level framework in `specified_framework_match`, without affecting the automatic evaluation (use `list_messaging_frameworks` to find available IDs). **Do NOT call** for content without a target persona, or for content with no marketing intent. **No relevant organization-level persona** (`list_personas` empty, or nothing matching the content's audience) → this tool is unavailable; don't invent a handle or stall. Assess the draft against `query_market_research` on its apparent audience, qualitatively — never a 0–100 figure this tool didn't produce — and offer persona setup as the upgrade that unlocks scoring. **Presentation:** Always render the dimensional scores as a visual bar chart or table before the written summary. Do not present results as a wall of text.

Returns information about who is authenticated — either a specific user or an org-level API key connection.