Analytics

Deterministic, database-grounded analytics over the sponsorship deal graph. Use these instead of free-text ask_graph for any counting, ranking, or benchmarking question.

---

aggregate_deals

Count sponsorship deals grouped by a facet — the deterministic answer to "how many / top N / by category|sport|league / benchmark" questions.

Parameters

Name	Type	Required	Description
group_by	string	Yes	One of industry, sub_sector, sport, league, country, deal_type, sponsor_tier
split_by	string	No	Secondary facet (same value set) — nests a count under each group
filters	object	No	{ industry, sub_sector, sport, league, country, deal_type:[], status:[], min_amount, since, target_label }. Values use exact English taxonomy (e.g. financial_services).
metrics	string[]	No	Subset of count, amount_avg, amount_median, amount_sum (amount metrics are EUR/USD/GBP only; not combinable with split_by)
top_n	number	No	Max groups (default 10, max 50)
include_examples	number	No	Flagship deals per group (max 5)

Returns

{ group_by, split_by, filters, groups: [{ key, count, splits?, amount?, examples? }], total_deals_considered }.

Example prompts

"Top 10 categories by deal volume over the last 12 months, split by sport" "How many sponsorship deals per sport for fintech brands?" "Kit/jersey sponsorship deals among La Liga clubs"

Note: deal_type distinguishes jersey, naming, branding, etc. — there is no dedicated sleeve value, so "sleeve" resolves to jersey/kit deals.

---

database_search

Faceted AND/OR entity filter over the graph. Two uses:

1. Build shortlists — filter brands/rightholders by industry, sub-sector, sponsored sport/league/country, and more.
2. Bulk-export many brands — one call returns up to 100 complete nodes, the source of truth behind the app's CSV export and the CRM sync. (For a SINGLE brand, get_entity already returns the full node + its sponsors.)

Parameters

Name	Type	Required	Description
groups	object[]	Yes	[{ operator: "AND"|"OR", conditions: [{ entity_type, property, operator, value }] }]. Condition operator ∈ eq, contains, in, gte, lte, range, not_eq, exists, not_exists.
operator	string	No	How groups combine (AND/OR)
limit	number	No	Max nodes (default 100, max 100)
offset	number	No	Page offset

Returns

{ nodes: [{ id, label, type, properties }], total }. For a brand, properties contains its full firmographic profile.

Every field of a brand

Filter on name to pull one brand's whole node:

{
  "groups": [
    { "operator": "AND", "conditions": [
      { "entity_type": "brand", "property": "name", "operator": "eq", "value": "Qonto" }
    ]}
  ]
}

The returned properties object includes (when known):

• Identity — name, aliases, website, website_domain, linkedin_url, description
• Classification — industry, sub_sector, keywords, business_models, brand_positioning, geographic_scope, ownership_type, group (parent)
• Size & financials — revenue, revenue_eur, revenue_range, employees, founding_year
• HQ — siege_social_pays (country), siege_social_etat (state/region), siege_social_ville (city), address, hq_continent, hq_region
• Geographic footprint — operating_countries, operating_regions, operating_cities, operating_continents, brand_acts_in
• Audience — target_audience_income_level, target_audience_key_interests
• Funding (VC-backed brands) — funding_rounds_count, funding_last_round_type, funding_last_round_amount_eur, funding_last_round_date, funding_first_round_date, growth_stage

For a single brand, get_entity returns the same full node (plus its sponsors). Use database_search when you need many brands at once or a filtered shortlist.

Multiple brands at once (bulk export)

A single call returns up to 100 full nodes — the same data the app's CSV export produces for a list. total reports the full match count; page with offset (0, 100, 200…) for larger sets.

• By filter — every brand matching a facet, each with its full profile:

  {"groups":[{"operator":"AND","conditions":[
    {"entity_type":"brand","property":"sub_sector","operator":"eq","value":"fintech"}]}],
   "limit":100, "offset":0}

• By an explicit list of names — use the in operator:

  {"groups":[{"operator":"AND","conditions":[
    {"entity_type":"brand","property":"name","operator":"in","value":["Revolut","Qonto","N26"]}]}]}

To export your saved list specifically, get_pipeline returns every saved entity in the workspace with its stored entityData — the MCP equivalent of the My Lists CSV export.

---

See also brand_intel_benchmark (avg/median/count of deal amounts by sport & league) and brand_intel_opportunities (under-served sport segments).

---

get_distinct_values

List the distinct values a property takes — powers typeahead and filter building.

Parameters

Name	Type	Required	Description
node_type	string	Yes	e.g. "Brand", "RightHolder"
property_name	string	Yes	Property to enumerate (e.g. "industry", "country")
prefix	string	No	Only values starting with this prefix
limit	number	No	Max values (default 20)

Example prompt

"What industries exist in the brand database?"

---

get_value_counts

Count how many nodes have a given property value.

Parameters

Name	Type	Required	Description
property_name	string	Yes	Property to match
value	string	Yes	Value to count
node_types	string[]	Yes	Which node types to count across

Example prompt

"How many brands are in the Retail industry?"