Intelligence

Due diligence reports, sponsorship matching scores, web traffic analytics, and workspace memory — the research toolkit.

`get_due_diligence`

Fetch a cached due diligence report for an entity. Does not trigger generation if no report exists yet.

Parameters

Name	Type	Required	Description
`entity_type`	`string`	Yes	`"brand"`, `"rightholder"`, `"athlete"`, `"event"`, or `"organization"`
`entity_name`	`string`	Yes	Entity's canonical name

Returns

Cached report with cards (leadership, financials, brand image, recent deals, sponsoring history, etc.) and citations, or null if not cached.

Example prompt

"Show me the cached due diligence on Red Bull"

`generate_dd`

Generate (or retrieve cached) a due diligence report using Claude Opus + web search.

Timing

Takes 15–30 seconds when generating. Returns the cached version immediately if already computed (unless force_refresh: true).

Parameters

Name	Type	Required	Description
`entity_type`	`string`	Yes	`"brand"`, `"rightholder"`, `"athlete"`, `"event"`, or `"organization"`
`entity_name`	`string`	Yes	Entity's canonical name
`language`	`string`	No	Report language: `"en"`, `"fr"`, `"es"`, `"it"`, `"pt"`, `"de"` (default: `"en"`)
`force_refresh`	`boolean`	No	Bypass cache and regenerate (default: `false`)

Returns

Full report: { report: { entity_name, entity_type, language, generated_at, cards: [...], from_cache } }.

Each card has key, title, data, citations, confidence, and optional score.

Example prompt

"Generate a due diligence on Adidas in French"

`generate_matching`

Generate sponsorship matching scores and sales attack angles between an entity and your workspace.

Timing

Takes 5–35 seconds total: fetches or generates the DD report first, then runs the matching LLM call.

Parameters

Name	Type	Required	Description
`entity_type`	`string`	Yes	`"brand"`, `"rightholder"`, `"athlete"`, etc.
`entity_name`	`string`	Yes	Entity's canonical name
`language`	`string`	No	Output language (default: `"fr"`)

Returns

{
  "report": {
    "scores": [
      {"key": "brand_alignment", "label": "Brand alignment", "score": 82, "justification": "...", "color": "green"},
      {"key": "audience_match", "label": "Audience match", "score": 74, "justification": "...", "color": "green"},
      {"key": "media_value", "label": "Media value", "score": 61, "justification": "...", "color": "orange"},
      {"key": "deal_viability", "label": "Deal viability", "score": 55, "justification": "...", "color": "orange"}
    ],
    "summary": "Strong brand alignment...",
    "angles": [
      {
        "dimension": "brand_alignment",
        "title": "Premium positioning synergy",
        "hook": "Both Adidas and [your entity] occupy the performance-premium tier...",
        "arguments": ["Shared target demographic 18–35...", "Co-visibility on European stadiums..."]
      }
    ]
  }
}

Example prompt

"Generate the matching between Adidas and my workspace, in English"

`get_matching_cached`

Read pre-computed matching scores for a brand that appears in your hot brands list.

Parameters

Name	Type	Required	Description
`entity_type`	`string`	Yes	Entity type
`entity_name`	`string`	Yes	Entity name

Returns

Same structure as generate_matching, or null if not yet computed.

`generate_memory`

Generate workspace memory fields by researching your entity on LinkedIn and the web. These fields improve generate_matching accuracy.

Timing

Takes 8–15 seconds (Claude Opus + web search).

Parameters

Name	Type	Required	Description
`workspace_name`	`string`	Yes	Your workspace / organization name
`linkedin_url`	`string`	Yes	LinkedIn URL of your organization
`language`	`string`	No	Output language (default: `"en"`)

Returns

{
  "institutionInfo": "Founded in 1998, [org] is a...",
  "fanbaseInfo": "Core fanbase: 18–35, predominantly male...",
  "sponsoringStrategy": "Current sponsors include...",
  "digitalEngagement": "2.1M followers on Instagram...",
  "socialLinkedin": "https://linkedin.com/company/...",
  "socialInstagram": "https://instagram.com/..."
}

`get_brand_traffic`

Web traffic analytics for a brand: monthly visit trend, source breakdown, geographic distribution, ad spend estimate.

Parameters

Name	Type	Required	Description
`brand_name`	`string`	Yes	Brand name
`domain`	`string`	No	Brand's primary domain (improves accuracy)

Returns

Traffic payload with visits (monthly series), sources (organic/paid/social/direct percentages), countries (top 5), ad_spend_estimate. Cached for 7 days.

Example prompt

"What is the web traffic of Nike.com?"

`save_entity_url`

Save a website URL to an entity's node in the graph database. Useful for entities missing a website — improves DD quality and traffic analytics.

Parameters

Name	Type	Required	Description
`entity_name`	`string`	Yes	Entity name
`entity_type`	`string`	Yes	Entity type
`url`	`string`	Yes	Website URL (e.g. `"https://www.entity.com"`)
`entity_id`	`string`	No	Entity ID (improves accuracy)

Returns

{ "success": true } or an error.

`generate_attack_angles`

Generate sales angles and a draft outreach email for a target, from its due diligence and your workspace memory. Runs a paid LLM pass, so it previews and asks for confirmation first.

Parameters

Name	Type	Required	Description
`entity_name`	`string`	Yes	Target entity
`entity_type`	`string`	Yes	Entity type
`dd_report`	`object`	Yes	The due-diligence report (from `get_due_diligence`)
`workspace_memory`	`object`	Yes	Your workspace memory profile
`workspace_name`	`string`	Yes	Your workspace name
`language`	`string`	No	Output language (default `"fr"`)

Example prompt

"Give me angles and a draft email to approach Adidas"

`generate_pitch`

Generate a commercial pitch for a target from its due diligence and matching report. Runs a paid LLM pass, so it previews and asks for confirmation first.

Parameters

Name	Type	Required	Description
`entity_name`	`string`	Yes	Target entity
`entity_type`	`string`	Yes	Entity type
`dd_report`	`object`	Yes	The due-diligence report
`matching_report`	`object`	Yes	The matching report (from `generate_matching`)
`workspace_memory`	`object`	Yes	Your workspace memory profile
`workspace_name`	`string`	Yes	Your workspace name
`language`	`string`	No	Output language (default `"fr"`)

Example prompt

"Write a pitch for Red Bull based on my matching"

Inline generation (context → save)

For due diligence, matching, and workspace memory, you can generate the report yourself inline (with your own web search) instead of paying for the backend LLM pass. The pattern is always: call the get_*_context tool to receive the research prompt + field schema, write the content, then call the matching save_* tool to persist it to the same store the app reads. These context tools are fast and do not call the backend LLM.

`get_memory_context`

Get the workspace-memory research prompt + field schema so you generate the memory inline, then call save_workspace_memory.

Name	Type	Required	Description
`workspace_name`	`string`	Yes	Workspace name
`linkedin_url`	`string`	No	Workspace LinkedIn URL
`language`	`string`	No	Default `en`

Returns { system_prompt, user_prompt, fields }.

`save_workspace_memory`

Persist a workspace-memory profile you generated. Fields include socialLinkedin, socialTwitter, socialInstagram, socialFacebook, socialTiktok, socialYoutube, socialWebsite, institutionInfo, fanbaseInfo, sponsoringStrategy, digitalEngagement (all optional strings).

`get_dd_context`

Get the entity context + due-diligence card structure + writing instructions so you generate each card (paragraph ≤ ~40 words + score) inline, then call save_due_diligence. Faster than generate_dd.

Name	Type	Required	Description
`entity_type`	`string`	Yes	e.g. `"brand"`
`entity_name`	`string`	Yes	Entity name
`language`	`string`	No	Default `en`

`save_due_diligence`

Persist a DD report you generated inline. cards = [{ key, content, score, citations }].

Name	Type	Required	Description
`entity_type`	`string`	Yes	Entity type
`entity_name`	`string`	Yes	Entity name
`cards`	`object[]`	No	The generated DD cards
`entity_id`	`string`	No	Entity id
`language`	`string`	No	Default `en`

`get_matching_context`

Get the DD report + workspace memory + matching instructions (6 dimensions, color thresholds, angles schema) so you compute fit scores + attack angles inline, then call save_matching.

Name	Type	Required	Description
`entity_type`	`string`	Yes	Entity type
`entity_name`	`string`	Yes	Entity name
`language`	`string`	No	Default `fr`

`save_matching`

Persist a matching report you generated inline (scores + summary + angles) so it surfaces in the app's matching UI.

Name	Type	Required	Description
`entity_type`	`string`	Yes	Entity type
`entity_name`	`string`	Yes	Entity name
`report`	`object`	No	The generated matching report
`language`	`string`	No	Default `fr`

get_due_diligence​

Parameters​

Returns​

Example prompt​

generate_dd​

Parameters​

Returns​

Example prompt​

generate_matching​

Parameters​

Returns​

Example prompt​

get_matching_cached​

Parameters​

Returns​

generate_memory​

Parameters​

Returns​

get_brand_traffic​

Parameters​

Returns​

Example prompt​

save_entity_url​

Parameters​

Returns​

generate_attack_angles​

Parameters​

Example prompt​

generate_pitch​

Parameters​

Example prompt​

Inline generation (context → save)​

get_memory_context​

save_workspace_memory​

get_dd_context​

save_due_diligence​

get_matching_context​

save_matching​

`get_due_diligence`

Parameters

Returns

Example prompt

`generate_dd`

Parameters

Returns

Example prompt

`generate_matching`

Parameters

Returns

Example prompt

`get_matching_cached`

Parameters

Returns

`generate_memory`

Parameters

Returns

`get_brand_traffic`

Parameters

Returns

Example prompt

`save_entity_url`

Parameters

Returns

`generate_attack_angles`

Parameters

Example prompt

`generate_pitch`

Parameters

Example prompt

Inline generation (context → save)

`get_memory_context`

`save_workspace_memory`

`get_dd_context`

`save_due_diligence`

`get_matching_context`

`save_matching`