Evidence vault for film studios that captures pre-production AI prompts and storyboards before copyright discovery hits.

1. Discovery scope is moving beyond consumer-facing releases, so internal AI workflow evidence now matters in court.
2. Storyboarding and ideation were named as relevant internal uses, making pre-production teams the first urgent capture wedge.
3. A litigant can now demand all prompts and outputs, which breaks teams that still treat prompt history as disposable creative scratch space.
4. A precedent from this dispute could spill into cases touching OpenAI, Anthropic, and Stability AI, so studios need a vendor-neutral record layer.
5. Potential statutory damages up to 150,000 dollars per infringed work justify software spend on evidence retention before one case balloons.

Catalyst. Midjourney's push to force disclosure of internal storyboarding, ideation, and full prompt and output histories makes backstage AI usage an immediate litigation and governance problem for studios.

The product connects to sanctioned Midjourney workspaces, creative review channels, and asset libraries, then stamps each prompt and output with project, user, timing, and downstream-use metadata. It separates purely internal ideation from assets approved for external use, so counsel can answer narrower questions without dumping an entire studio's experimentation history. When a dispute or preservation notice lands, the system freezes relevant sessions, maps output lineage to edits and exports, and opens a privilege- review workspace instead of a raw file scramble. Policy rules can require project tagging, retention windows, and manager approval for shared AI workspaces, giving studios defensible process before a court ever asks. Over time, the vault becomes the canonical evidence layer studios, outside counsel, and insurers trust when evaluating AI-assisted productions.

What's different. Most adjacent vendors focus on what ships publicly, on asset licensing, or on generic post-hoc e-discovery. This company owns the missing layer between internal creative experimentation and legal response: continuously captured, project-scoped AI workflow evidence with privilege and retention controls. Because it maps lineage across prompts, outputs, edits, and productions over time, it compounds into a studio-specific record generic storage, DLP, or legal-hold tools cannot recreate.

Jobs to be done

flowchart LR
  Buyer[Studio legal and creative ops] --> Pain[Cannot reconstruct internal AI storyboards under discovery]
  Pain --> Product[Creative AI evidence vault]
  Product --> Outcome[Faster defensible responses without halting pre-production]

Executive takeaways

• The wedge is upstream preservation of creative AI exhaust: discovery is already reaching internal storyboarding, ideation, prompts, and outputs, not only shipped assets.
• Budget adjacency is strong because legal hold, e-discovery, and creative workflow systems already absorb the cost of manual preservation and review.
• The first credible buyer is the Deputy GC for IP or litigation with a creative-technology sponsor, not the art department alone.
• The strongest substitutes are generic holds, storage exports, and provenance labels, so the startup must win on cross-tool capture plus matter-ready packaging.
• Execution risk is mainly technical coverage: Midjourney activity spans privacy modes and Discord contexts while native logs are time-limited or fragmented.

Market definition

Software that preserves and packages internal generative-media evidence for film and TV production teams: prompts, outputs, edits, approvals, release status, and legal-hold context across creative AI and collaboration tools.

Customer and buyer

Primary users are AI creative-ops leaders, production technology managers, and litigation-support staff who need to freeze and reconstruct creative workflows by production. The economic buyer is usually the Deputy GC for IP or litigation, with the Head of Creative Technology as the operational sponsor because the problem crosses legal risk, workflow governance, and pre-production continuity.

Buying triggers

• A discovery request or preservation notice now reaches internal AI prompts, outputs, and storyboarding work, forcing studios to identify custodians and ESI before public release is even at issue. [1][2][4][13][14][17][19]
• A sanctioned rollout of Runway-style or Midjourney-based pre-production workflows creates pressure to separate internal experimentation from release-ready assets and to keep project-level histories. [6][7][8][32][33][34]
• Union, copyright, and AI-governance scrutiny makes buyers ask what AI-generated material was provided to writers, how outputs were created, and whether training or rights questions can be answered later. [9][10][11][12][20][21]

Willingness to pay

Willingness to pay is credible because the problem already burns budget in adjacent systems: legal holds, e-discovery review, collaboration retention, and outside-counsel collection. The startup is not inventing a new line item so much as collapsing manual matter-response work into a cross-tool evidence layer before sanctions, statutory damages, or expensive over-collection arrive. [16][18][19][26][27][29][30][40]

Category dynamics

Growth signal 10.4% CAGR (adjacent eDiscovery software, 2025-2030)

Validation signals

• Midjourney is explicitly seeking internal AI business plans, storyboarding uses, and full prompt or output history in discovery.
• Hollywood studios and networks are formalizing AI vendor relationships for pre-production, development, and marketing workflows.
• The collaboration stack already supports holds and preservation primitives, so buyers can imagine the workflow without a wholly new infrastructure religion.
• Adjacent eDiscovery software is growing as AI features move work into software, signaling buyer willingness to pay for workflow compression.

Regulatory & technical constraints

• Rules 26 and 34 turn prompts, outputs, images, and related metadata into discoverable ESI when relevant and within control; Rule 37 raises sanctions risk if preservation fails.
• Copyright statutory damages can scale per work, magnifying the cost of sloppy evidence boundaries.
• The Copyright Office is separately examining digital replicas, output copyrightability, and training, so studio AI policies will keep moving.
• WGA rules require disclosure when AI-generated material is provided to writers, creating an operational need for provenance and workflow tagging.
• Discord audit logs are retained for 45 days and Slack legal holds have scope limits, so native collaboration logs alone are not enough as a long-horizon system of record.

Competition is fragmented across e-discovery, collaboration retention, creative provenance, and enterprise AI governance. No incumbent owns the narrow system-of-record layer that continuously captures prompts, outputs, human edits, and release status by production before a matter exists.

Why incumbents do not win by default

• E-discovery platforms. Relativity-class and Everlaw-class platforms are trusted once a matter exists, but they are optimized for legal hold, collection, and review after the preservation scramble has already begun.
• Collaboration and storage systems. Slack, Box, and Discord can preserve pieces of the record, yet they remain fragmented systems with scope limits, short-lived logs, and weak production-level release-state context.
• Creative provenance and review stack. Adobe, C2PA, and Frame.io improve asset provenance and review history, but they do not by default own counsel-ready prompt history, legal holds, or cross-tool matter packets.
• AI-governance control planes. Credo-class governance tools are strong at registries, policies, and enterprise AI risk workflows, but they are not built around studio-specific creative exhaust and production evidence.
• Native AI vendors and manual collections. Midjourney privacy settings and native collaboration logs help at the margin, but vendor-specific controls still leave legal teams reconstructing the full chain across tools.

Film and episodic studios are moving Midjourney-style ideation into shared pre-production workflows just as discovery requests are reaching internal prompts, outputs, storyboards, and related metadata. The company sells a vendor-neutral evidence vault that continuously captures sanctioned AI creative activity by production, then generates matter-scoped legal-hold and discovery packets before outside counsel has to reconstruct history from chats and exports. The initial beachhead is deliberately narrow: Los Angeles or U.S.-centric studios and streamers with centralized storyboard or concept-art teams and at least one shared Midjourney workflow across multiple active productions. That wedge is credible because the economic buyer already spends on preservation, e-discovery, and outside-counsel collection, so the product can reallocate existing risk and operations budget rather than invent a speculative AI budget line. The go-to-market is one coherent system: sell a paid legal-hold-readiness pilot to a Deputy GC triggered by a preservation notice or AI policy rollout, sponsor it with the Head of Creative Technology, price it by active productions and connected workspaces, and expand only after the first matter packet proves faster and narrower than manual collection. The strongest moat is a production-level evidence graph linking prompts, outputs, edits, release state, and prior matter-response patterns across tools that generic holds, DAMs, and provenance labels do not assemble. Two diligence gaps remain open: how much relevant workflow actually stays inside sanctioned studio-owned workspaces, and whether legal or creative-technology leaders own the first budget in practice. If the first two pilots cannot reach useful coverage and convert to annual contracts at software-like gross margins, the opportunity is a services niche rather than a venture-scale platform.

Problem

• Discovery and preservation now reach internal AI storyboarding, ideation, prompts, outputs, and related metadata before any asset is publicly released, so studios can no longer treat pre-production AI usage as disposable scratch work.
• The record is fragmented across Midjourney workspaces, Discord or Slack channels, Box folders, asset review tools, and manual exports, forcing legal teams to over-collect or reconstruct creative history by hand under time pressure.
• Generic substitutes such as legal holds, storage retention, provenance labels, and outside-counsel collections preserve pieces of the story but do not create a production-scoped prompt-to-output-to-edit evidence packet with release-state context.

Solution

• Connect sanctioned Midjourney workspaces plus adjacent Discord, Slack, Box, and review surfaces to continuously capture prompts, outputs, user identity, timestamps, project tags, and downstream asset references by production.
• Apply release-state labels, retention rules, and matter-scoped legal holds so counsel can freeze the right sessions and generate narrow discovery packets without dumping an entire studio's experimentation history.
• Provide a privilege-review workspace and outside-counsel access model that turns captured creative exhaust into a repeatable response workflow rather than a one-off forensic scramble.

Why we win

• The company sits upstream of Relativity-, Everlaw-, Adobe-, and collaboration-stack alternatives by capturing evidence before a matter exists instead of after custodians and files must already be identified.
• Vendor-neutral, production-level lineage across prompts, outputs, edits, approvals, and release state is the missing system of record that neither native AI vendors nor provenance tools currently own.
• Internal IT or creative-ops teams can script exports, but they do not want to maintain evolving connector coverage, privilege workflows, and matter-ready packet design across legal and creative systems.
• Starting with one urgent storyboard-evidence wedge creates referenceable proof faster than a broad media-governance platform and compounds into a studio-specific evidence graph that becomes harder to replace over time.

Milestones

flowchart LR
  Wedge[Studio AI evidence wedge] --> MVP[Midjourney plus collaboration capture]
  MVP --> Proof[48-hour matter packets with 80% plus sanctioned coverage]
  Proof --> Expansion[More productions tools and adjacent media buyers]

Founding team

Experiment roadmap

Risk assessment

What must be true

• At least 10 of the ~25 U.S. beachhead accounts already run shared creative-AI workflows across multiple active productions.
• Deputy GCs will sign a paid pilot before a live sanctions event if the trigger is a preservation notice or policy rollout.
• Sanctioned-workspace capture plus collaboration and storage ingest can cover at least 80% of relevant storyboard activity in the first two customers.
• A production-scoped packet can be generated in 48 hours or less and materially reduce outside-counsel collection effort versus manual exports.
• Relativity, Everlaw, Adobe, and collaboration stacks will not close the upstream capture gap before the startup lands five referenceable accounts.

Open diligence questions

• Which named studios already run studio-owned Midjourney or Runway workspaces across multiple productions?
• What is the current time and outside-counsel cost to respond to one AI preservation request by production?
• Who controls the first budget and procurement path on this joint legal and creative-tech problem?
• What share of relevant workflow sits outside sanctioned workspaces or disappears behind privacy modes?
• Do counsel and insurers care enough about release-state lineage to prefer this over generic holds plus manual review?

Model sanity

• Revenue engine. Base revenue comes from growing paying studio accounts from 3 at Y1 exit to 9 by Q4Y3 while blended account value rises toward the researched ~$300K ACV through scope expansion.
• Must go right. The capture-and-packet workflow must stay template-driven enough that 3 engineers and 2 solutions hires can support 6 customers by Q4Y2 without gross margin stalling below 70 percent.
• Model breaks if. If pilot-to-annual conversion slips toward 120-150 days or account value lands below the stated pricing band, the downside case drives the cash floor toward roughly $0.1M before the referral channel is proven.
• Next-round proof. The next financing story is 8 paying accounts by Q2Y3 with a second creative surface live, counsel-sourced referrals contributing pipeline, and quarterly EBITDA close to breakeven.

Scenarios

Sensitivity

flowchart LR
  Trigger[Preservation notice or AI policy trigger] --> Pilot[Paid readiness pilot]
  Pilot --> Rollout[Annual production rollout]
  Rollout --> Expansion[Hold seats and second-surface expansion]
  Expansion --> Revenue[Revenue]
  Revenue --> GrossProfit[Gross profit]
  GrossProfit --> Cash[Cash and runway]

Flags: Y1 customersEop includes paid pilots as well as annual subscriptions, so recurring ARR at the first year-end is lower than the simple customer count suggests. · The Y3 revenue path requires the company to reach the upper half of the stated $250K-$350K annual pricing band through hold, retention, and second-surface expansion modules. · Gross margin only clears the 70 percent target if deployments become template-driven; if coverage audits stay custom, the business can slide toward a services profile. · The near-term beachhead is only about 25 U.S.-centric accounts, so over-hiring GTM ahead of referrals would quickly damage CAC and runway. · Cash is modeled as EBITDA with immaterial working-capital timing; implementation prepayments or deferred revenue could move actual cash somewhat earlier or later.

• Capture blind spots. Creative teams may use unsanctioned tools, side chats, or local workflows that the product cannot observe on day one. Mitigation: Start with sanctioned shared workspaces and chat or asset ingest, then expand coverage through policy-backed connectors and historical imports.
• Split-budget sales motion. Legal feels the discovery pain, but creative-tech or IT may own the tools and budget, slowing deal cycles. Mitigation: Sell one quantified matter-response workflow with joint legal and creative-ops champions plus outside-counsel validation.
• Beachhead concentration. The number of Hollywood studios with mature internal AI pre-production workflows may be smaller and slower-moving than expected. Mitigation: Use flagship studio design partners to expand quickly into streamers, game studios, agencies, and insurer-driven compliance reviews.