CLIP™ Executive Deck Pitch Case Study Workshop Architecture Technical PRD
Product Storybook · April 2026

CLIP™

Contractual Licensing Intelligence Platform
From 80-page legal contracts to live revenue intelligence — the complete story.

Module 1 — Contract Intelligence
Module 2 — Product Metadata Intelligence
CLIP Core™ · Insights™ · Forecast™
Chapter 1 — The Problem

A New Lead Joins the Team

Meet Alex. Senior sales planner, just joined. First task: figure out which titles can be licensed to a streaming platform under a new output deal.

👩‍💼
Alex — Senior Sales Planner, Day 1
"My manager handed me a 94-page PDF. Apparently it's the contract. I need to find which of our 12,000 titles are eligible — and which window each one falls into. There are territorial exclusions, date offsets, AND/OR conditions... and three amendments. Where do I even start?"
😰
The Reality Without CLIP™
"Two weeks later... I've read the contract four times. I'm still not sure if 'window commences 100 days after theatrical release in USA' means the initial, general or limited release date. And I have no idea which of our titles have a theatrical release date populated at all."
✨
With CLIP™ — Day 1, Same Contract
"Upload the PDF. Module 1 extracts every rule. Module 2 checks our catalog and confirms which attributes are actually populated. In 20 minutes I'm looking at a dashboard showing 847 eligible titles, 234 conditional, and 412 forecast to open next quarter."
Module 1 — Contract Intelligence
Chapter 2 — The First Module

OCR Reads the Contract. LLM Understands It.

Module 1 is the pipeline that transforms a legal document into executable business rules. Four steps, fully automated.

📄
Step 1 — Ingest
PDF or Word uploaded. OCR converts every page to clean, structured text.
→
🧠
Step 2 — Extract
LLM reads the text and identifies eligibility clauses, date formulas, territory conditions, fallback logic.
→
📐
Step 3 — Structure
Legal language converted to versioned JSON rules with AND/OR logic, operators and territory filters.
→
✅
Step 4 — Review
Human reviewer inspects every rule in the UI. Corrects edge cases. Approves. Only then does execution begin.
👩‍💼
Alex — understanding Module 1
"So Module 1 is like having a very precise legal analyst read the contract and fill in a structured form. The form is JSON. And before it runs against any titles, I get to check the form and fix anything the AI misread. That's the trust mechanism."
Module 2 — Product Metadata Intelligence
Chapter 3 — The Second Module

The Catalog Knows What Attributes Are Real.

Module 2 reads the existing product catalog to discover which attributes actually exist and have data — then feeds that knowledge into Module 1 to keep the LLM grounded.

Without Module 2 — the hallucination problem
The LLM reads "theatrical release in USA" and might invent a field name like theatricalDate_US — which doesn't exist in our system. The rule is extracted but it matches nothing when it runs against the catalog. Silent failure. No titles found. Revenue missed.
With Module 2 — grounded extraction
// Module 2 injects this reference list
// into Module 1 before LLM runs:
 
knownAttributes: [
  "initialTheatricalReleaseDate",
  "generalTheatricalReleaseDate",
  "standardEstReleaseDate",
  "standardPhysicalReleaseDate"
]
 
// LLM now maps "theatrical release"
// → initialTheatricalReleaseDate ✓
👩‍💼
Alex — understanding Module 2
"Module 2 is like giving the AI a cheat sheet before the exam. Instead of guessing field names, it picks from a validated list of attributes that we know exist in the catalog. It's also how the system knows which of our 12,000 titles actually have a US theatrical date populated — so we don't run rules against empty data."
Module 1 — Rule Extraction
Chapter 4 — Reading Legal Language

Legal Language → Executable Rules

The contract says things like "theatrical motion pictures released on 250 or fewer screens." CLIP™ turns that into AND/OR rule logic that runs against every title in seconds.

What the contract says
Qualifying titles shall include all theatrical motion pictures that:

✦ Had a theatrical release date in the USA AND
✦ Have reported box office data in the USA AND
✦ Were released on 250 or fewer screens AND
✦ Either had P&A expense of $2.5M or less
  OR were produced in Color format
What CLIP™ extracts
"condition": "AND", "rules": [
  { attr: theatricalReleaseDate
    op: NOT_NULL, territory: USA },
  { attr: boxOffice
    op: NOT_NULL, territory: USA },
  { attr: numberOfScreens
    op: ≤ 250 },
  // nested OR block:
  { "condition": "OR", "rules": [
    { pnaExpense ≤ 2,500,000 },
    { colorBw = "Color" }
  ]}
]
ANDAll conditions must be true for a title to qualify
theatricalReleaseDate≠NULL📍 USA
boxOffice≠NULL📍 USA
numberOfScreens≤250
ORAt least one of these
pnaExpense≤$2,500,000
colorBw="Color"
Module 1 — Date Formulas
Chapter 5 — When Does the Window Open?

Date Offsets. Earlier Of. No Later Than.

Contracts don't say "January 1st." They say things like "the earlier of 100 days after theatrical in USA or 90 days after EST in Germany." CLIP™ makes that computable.

What the contract says
"Window commences on the earlier of:

✦ 100 days after the Initial, General, or Limited Theatrical release in the United States

✦ 100 days after the Standard EST, Standard VOD, Early EST, PEST, PVOD, or Physical (LVR) release in the United States

✦ As a final fallback: 100 days after the US Theatrical window start date"
What CLIP™ extracts
"conditionId": 5 // EARLIER OF
"formulas": [
  { attr: initialTheatricalReleaseDate
    +100 DAY, territory: USA(235) },
  { attr: generalTheatricalReleaseDate
    +100 DAY, territory: USA(235) },
  { attr: standardEstReleaseDate
    +100 DAY, territory: USA(235) },
  // ... more formulas
]
"fallbackRules": [
  { priority: 3, sourceWindowDef: 386 + 100 DAY }
]
Real Production Data — CEE: Baltics
Chapter 6 — A Real Window Pattern

CEE: Baltics GP1 — The Three-Priority Cascade

This is real data extracted from the CLIP™ system. The window for CEE Baltic territories uses a three-priority fallback chain — a common real-world pattern.

CEE: Baltics GP1
12 months
P1
Earlier of: 100 days after General Theatrical Date, Initial Theatrical Date, Limited Theatrical Date in United States of America
P2
Earlier of: 100 days after Early EST, PEST, PVOD, Standard EST, Standard Physical (LVR), Standard VOD in United States of America
P3
100 days after Theatrical US Start date (references window definition 386 directly)
1
Priority 1 — Theatrical Dates in USA
System tries each of these dates + 100 days and takes the earliest result. If any exist, this resolves the window start date.
initialTheatricalReleaseDate
generalTheatricalReleaseDate
limitedTheatricalReleaseDate
2
Priority 2 — Home Entertainment Dates in USA
No theatrical dates? Fall back to digital and physical release dates. Takes the earliest + 100 days.
standardPhysicalReleaseDate
standardEstReleaseDate
earlyEstReleaseDate
premiumEstReleaseDate
premiumVodReleaseDate
3
Priority 3 — US Theatrical Window Start (final fallback)
No release dates at all? References the US Theatrical window definition start date directly and adds 100 days. Every title gets a window — no silent failures.
windowDefinitionId: 386 + 100 DAY
Module 1 — Dynamic References
Chapter 7 — The Home Office Problem

Some Rules Change Per Title. That's Normal.

Contracts sometimes say "fall back to the release date in the Home Office territory." The problem: Home Office is different for every movie depending on where it was made.

👩‍💼
Alex — encountering the Home Office clause
"I found a clause that says 'if no other release date is available, use the Home Office release date.' But Home Office isn't a fixed territory — it's the country where the movie was produced. A French film's Home Office is France. A Korean film's is Korea. How does the system handle that?"
Contract Rule
  • Fallback: release date in Home Office territory
  • Home Office = title's country of origin
  • Placeholder: $HOME_OFFICE
→
CLIP™ Resolution at Runtime
  • Reads title's countryOfOriginList
  • Maps country → territoryId
  • Fetches release date for that territory
  • French film → France → FR theatrical date
  • Korean film → Korea → KR theatrical date
Plain English — what it means
The rule is defined once but resolves differently for each title. A French co-production will use its French theatrical release date. A US film will use its US date. The system looks up each title's country of origin and fetches the right date automatically — no manual lookup needed.
How CLIP™ stores it
"fallbackFormula": {
  "formulaAttribute": "earliestRelDate",
  "filterByAttributes": [{
    "attribute": "territoryId",
    "values": ["$HOME_OFFICE"],
    "dynamicReference": {
      "sourceAttribute":
       "countryOfOriginList"
    }
  }]
}
Module 1 + Module 2 — Versioning
Chapter 8 — Contracts Change. So Do Patterns.

Published Today. Draft for Tomorrow.

The real CLIP™ data shows two versions of the same pattern existing side-by-side. The Published version is live. The Draft version is the pending update, waiting for its effective date.

✅ Published — Active Now
CEE: Baltics GP1
Pattern ID: 237
Effective Date: 2026-01-29

Applies to titles with US Theatrical start on or after 2025-12-22.

Currently running against the catalog. Producing window dates for eligible titles today.
🟡 Draft — Pending Update
CEE: Baltics GP1
Pattern ID: 283
Effective Date: 2026-02-24

Will apply to titles with US Theatrical start on or after 2026-09-23.

Same rule logic as Published. Becomes the active version automatically when its effective date is reached.
👩‍💼
Alex — on versioning
"So when the contract is amended — say the window offset changes from 90 days to 100 days — the team creates a new Draft pattern with the new logic and the future effective date. The old Published pattern keeps running for existing titles. The new one kicks in automatically for new releases. No manual switchover. No risk of running the wrong rules."
Module 1 + Module 2 — Execution
Chapter 9 — Running the Rules

12,000 Titles. Rules Applied. Results in Seconds.

Once rules are approved in the UI, the rule engine runs them against the full title catalog. Every title gets evaluated and classified.

✅
Eligible
All eligibility conditions pass. Window is open right now. Ready to license immediately.
847
⏳
Conditional
Rules pass, but the window hasn't opened yet. Date formula calculated — waiting for the date to arrive.
234
🔮
Forecast
Will become eligible once a future release date condition is met. Pipeline visibility for next 12 months.
412
👩‍💼
Alex — on segmentation
"This is what I needed on Day 1. The 847 eligible titles I can go to market with right now. The 234 conditional ones I can plan deals around — I know exactly when their windows open. And the 412 forecast titles are my pipeline for the next year. No more guessing. No more re-reading the contract."
10K+
Titles evaluated per run
<60s
Full catalog execution
100%
Rules auditable
12 mo
Pipeline visibility
CLIP Core™ · Insights™ · Forecast™
Chapter 10 — The Platform

Three Tiers. One Platform. Grows With You.

CLIP™ is structured in three layers that build on each other. Start with Core — digitise your contracts. Add Insights — see what you can sell. Add Forecast — see what you will sell.

🏗️
CLIP Core™
Tier 1 — Contract Digitisation
✓ OCR + LLM ingestion
✓ Rule extraction
✓ Amendment versioning
✓ Master data mapping
✓ Human review UI
→
🔍
CLIP Insights™
Tier 2 — Eligibility Intelligence
✓ Everything in Core
✓ Rule engine on catalog
✓ Eligible / Conditional / Forecast
✓ Rights system integration
✓ Scenario modelling
→
📡
CLIP Forecast™
Tier 3 — Predictive Strategy
✓ Everything in Insights
✓ 12-month forecasting
✓ Amendment simulation
✓ Revenue modelling
✓ Pipeline dashboards

From Confusion
to Clarity

Alex's two-week manual grind becomes a 20-minute automated workflow. The same platform that helps a new lead get up to speed in minutes drives a 5.2× ROI and a 12-month revenue pipeline for the organisation.

Module 1 — Contract Intelligence
Module 2 — Product Metadata Intelligence
CLIP Core™ · Insights™ · Forecast™
5.2× First Year ROI
$6M+ Annual Value