Media & Entertainment AI agent · Consent
Digital-Replica & Performer-Consent AI Agent
Keep the register of digital replicas on a title, match each proposed use against the scope of the consent that was signed, and hold it there until a business-affairs executive approves.
- Typical delivery
- 4–6 weeks
- Deployment
- Your stack
- Exec approval
- Per-use consent
- After launch
- Agent Care
What this agent does
Assembles the consent record, never the authorisation
In
- 01
A replica is registered from production photography, a scan session or a VFX vendor manifest.
- 02
A capture is matched to the engagement, so a footage-derived asset is registered like a scanned one.
Reason
- 03
A new use is proposed, and it is read against the description of intended use in the signed consent.
- 04
An agreement version is pinned to the project — the 2026 SAG-AFTRA TV/Theatrical cycle runs to 30 June 2030.
- 05
A consent artefact is checked for bold-and-capitals form, a separate signature and a named project.
Decide
- 06
A gap is raised — no consent, a description too thin to reach the use, or a replica with no engagement.
- 07
A determination is routed to the named business-affairs executive, who decides whether the use proceeds.
Out
- 08
A consent does not lapse on death, so the record outlives the production — scope, not a term, is the limit.
- 09
A write happens only inside the approval boundaries agreed during implementation, never ahead of approval.
Product statement
The agent matches a use to the consent; a performer signs the consent, a business-affairs executive approves the use, and the union bargains.
Example workflow
One proposed use, register to approval
- 1Replica identifiedScan record, VFX vendor manifest, shot list or footage-derived asset inventory
- 2Consent assembledThe signed artefact, its use description, the notice served and the script version delivered
- 3Scope match proposedIn scope, out of scope or ambiguous, with the consent text cited and confidence
- 4Controls appliedArtefact-form checks, California and New York definition tests, clock checks and threshold
No human action required
Stages 1 to 4 run unaided, and no use is authorised at any of them — the agent is assembling, and the executive's lane opens at the confidence gate.
Goes to the business-affairs executive.
Adds a production counsel read first.
Executive approval
The use is held with its consent citation, its gaps and the confidence.
Approve · Amend · Send to counsel
Approved — cleared for use
- 6Project systems updatedOnly where write access and approval policy allow it
- 7Outcome evaluatedScope-call reversals, replicas missed at the register and consent findings raised after release
Amendments
A use found unconsented after release is re-determined and counted.
What should not run autonomously
Human approval stays in control
Outside the boundary — human approval required
8 items
- Signing a digital-replica consent.
- Deciding that a proposed use may go ahead.
- Serving union notice and bargaining a synthetic.
- Stating the articulable business reason for a scan.
Automation boundary
Agent acts unaided
- ✓Maintain the replica register from capture and vendor sources.
- ✓Match a proposed use to the consent's description of intended use.
- ✓Check the consent artefact against the formal requirements.
- ✓Run the clocks, raise the gaps and hold the use back.
Any write happens inside the boundaries agreed at implementation, never ahead of the approval.
- Warranting a consent chain to a distributor.
- Distributing a voice model of a named performer.
- Setting an ICDR's engagement terms.
- Changes to consent, scope or approval rules.
Example output
One proposed use, annotated
Everything the agent determines is attached to the consent it was read from.
Consent-match output · single use
Illustrative example- Production
- Series, season two
- Proposed use
- Dub of a scene into two territory versions
- Replica type
- Employment-based
- Consent cited
- Signed replica consent
- Confidence
- 92%
- Determination
- In scope — dubbing described
As received
Taken from the register and the signed consent on file — nothing on this side is written by the agent.
Consent evidence used
- Separate signature
- Use description
- Script version on file
Why this determination
The consent describes dubbing for this picture — the call the executive approves.
What the score decides
Below the configured threshold the use picks up a counsel read before it reaches the executive.
Action
ApproveAmendSend to counselValue
Where AI adds value
The same four claims, placed at the point in the workflow where each one applies.
Where the value lands
Value 01 – 04
Each registered use
From the replica register
Match against the consent
Read the signed consent's description of intended use, the notice evidence and the script version on file.
A replica need never be scanned
Footage-derived likenesses belong in the register beside scanned ones — the trigger is existence, not capture.
Answer what other agents defer
Our dubbing and marketing agents generate no voice or likeness; this one holds the record when a vendor does.
Build an evidence trail
The use, the consent it was matched to and the executive who approved stay on the project.
Integrations
Typical integrations
Five system groups connect to the same agent. Which of them are in scope is decided in discovery.
Casting and engagement
- Casting Networks
- Breakdown
- Deal memos
- Exhibit E
Production and VFX
- ShotGrid
- ftrack
- Vendor manifests
- shot lists
Business affairs
- RightsLogic
- FilmTrack
- DocuSign
- contract stores
Agent
Digital replica & consent
- Reads the register
- Matches to consent
- Holds for approval
Asset management
- MAM
- DAM libraries
- Iconik
- Dalet
Observability & evaluation
- OpenTelemetry
- Langfuse
- Supported monitoring/evaluation sources
Integration availability depends on the client's existing systems and API access.
Agent controls
Six layers between the model and the master
Six controls, the outermost first. What the whole set misses is written into the map below.
L6 · Outermost — last line of defence
Inward → L1 · closest to the model
- L1 – L2
- Test whether a use may stand
- L3
- Puts the release in an executive's hands
- L4 – L5
- Keep the use and the consent behind it
- L6
- Holds the use at unconsented when signals degrade
How Nestack evaluates it
Evaluate the consent workflow — not only the final determination.
Coverage runs the whole depth of the workflow, and every layer is cut by slice.
Surface — the use the company ships
- E1Final-output evaluationWas the scope call right against the consent as written?
- E2Step-level evaluationDid the agent use the right consent, script version and state test?
- E3Tool evaluationDid it read the correct replica asset and the correct project?
- E4Confidence calibrationDo low-confidence determinations actually attract more amendments?
- E5Slice evaluationHow does performance change across specific replica types?
- E6Business outcomeHow many uses needed an amendment or a correction after release?
Floor — the outcome the studio answers for
Failure modes
Where each failure originates in the agent
Seven failure modes, placed at the stage each one originates.
Agent lifecycle
Direction of processing →
01 · Retrieval
1 mode
- RY-03
Register built from scans
A footage-derived replica never enters the register.
Stage gathers
Scan records, vendor manifests and shot lists
02 · Reasoning
2 modes
- RY-04
Buried consent accepted
A signed deal memo is read as a replica consent.
- RY-06
Wrong script version
An alteration is cleared on a revision never seen.
Stage proposes
The scope call, the consent cited and confidence
03 · Tool / write
2 modes
- RY-02
Determination taken as approval
An assembly step stands where approval should be.
- RY-05
Transfer window lapses
A library sale runs the window unserved.
Stage writes
Only where write access and approval policy allow it
04 · Output
1 mode
- RY-01
Use outside consent scope
A use is cleared that the description does not reach.
Stage returns
The determination the executive signs against
05 · Change / Version
1 mode
- RY-07
Silent scope widening
A model or rule change loosens what counts as in scope.
Stage tracks
Model, prompt, consent rules and agreement version
- Sev-1 · a use ships with no valid consent
- Sev-2 · a wrong scope call reaches the file
- Sev-3 · source degrades, use routes to review
Affected slices
Consent coverage can hide one exposed cohort
A slate-wide consent-match rate can look acceptable while a small number of replica cohorts carry most of the unmatched uses. Nestack reports the unmatched-use rate by slice, not only in total.
Slice performance — reported separately, not only in aggregate
Illustrative example| Slice | Failure rate | Lift | Lift vs. threshold | Status |
|---|---|---|---|---|
| Footage-derived replicas | 9.4% | 3.8× | Review | |
| Off-card background engagements | 7.2% | 2.9× | Review | |
| Deceased-performer catalogue | 4.4% | 1.8× | Watch | |
| Scanned employment replicas | 1.6% | 0.7× | Normal |
Bar: unmatched-use-rate lift vs. scanned-replica baseline · scale 0–4.0× · tick marks the 2.0× review threshold
2 of 4 slices over threshold
Evidence-linked improvement
The loop closes on a consent case
A cycle closes when the unconsented use is a case the next release has to catch. That suite is what the next use proposed is measured against.
Improvement cycle · five stages
Switchback — the path turns at Improve and returns at Learn
- 01
Detect
Unmatched-use rate rises in a replica slice.
- 02
Diagnose
The performer who agreed to one picture and appeared in a second is traced to one cause.
- 03
Improve
The fix carries a number, and the uses that forced it travel with it.
- 04
Verify
Nothing goes out until every touched consent case has passed again.
- 05
Learn
It is kept for good, and the consent rules are amended in the same commit.
Learn → Detect
The return edge. The next detection runs against a suite one consent case longer.
Typical build scope
Twelve workstreams across six weeks
The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, sources, consent workflow, evaluation, integration, then production validation and handover.
Workstream
Week 1Week 2Week 3Week 4Week 5Week 6
- 01Consent workflow discovery and boundary definition.Week 1
- 02Register, casting and VFX source review.Week 1, Week 2
- 03Consent-artefact, scope and state-rule mapping.Week 1, Week 2
- 04Replica register ingestion and matching.Week 2
- 05Scope-match logic and consent binding.Week 2, Week 3
- 06Confidence scoring and gap routing.Week 3
- 07Executive approval workflow.Week 3, Week 4
- 08Asset and contract-system integration.Week 5
- 09Consent and scope cases.Week 4, Week 5
- 10Guardrails and approval controls.Week 4
- 11Project-trail instrumentation.Week 5, Week 6
- 12Deployment, documentation and Agent Care handover.Week 6
12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallel
Final scope and sequence confirmed in discovery
Engagement tiers
What each tier includes
Rows are the capabilities named in each tier's scope. Higher tiers include everything below them.
| Capability✓ in scope · — not at this tier | PilotOne title, one office | ProductionProduction business affairs | AdvancedMultiple slates / labels |
|---|---|---|---|
| Introduced at Pilot | |||
| Matching to your consents and rules | ✓ | ✓ | ✓ |
| Executive approval | ✓ | ✓ | ✓ |
| Scope-match baseline | ✓ | ✓ | ✓ |
| Introduced at Production | |||
| Reporting by production | — | ✓ | ✓ |
| Approval workflow in your systems | — | ✓ | ✓ |
| Approved write-back | — | ✓ | ✓ |
| Asset-management integration | — | ✓ | ✓ |
| Introduced at Advanced | |||
| Multi-territory catalogues | — | — | ✓ |
| Multi-stage legal approvals | — | — | ✓ |
| High replica volume | — | — | ✓ |
| Multi-agreement consent controls | — | — | ✓ |
| Build price | From $5,000 | From $8,000 | Custom quote |
- Final build price
- Confirmed after discovery based on integrations, workflow complexity, transaction volume, approval controls and deployment requirements.
- Separate from build
- Build pricing is separate from recurring Agent Care, which covers managed monitoring, evaluations, incidents and verified improvements after launch.
What we need from you
What you bring, and what we build with it
Each input maps to a piece of build scope and a week in the delivery timeline.
You bring
We build with it
01Your replica register and capture sources
→Replica register ingestion and source mapping
Week 102Representative signed consents
→Scope-match baseline and consent-artefact checks
Week 203Your approved consent forms and scope language
→Consent-artefact, scope and state-rule mapping
Week 104Access to relevant APIs, feeds or exports
→Register, casting and VFX assessment, then integration setup
Week 205Uses you would not want defended
→Scope cases and failure-mode testing
Week 406What no use may rest on
→Confidence scoring, gap routing, guardrails and approval controls
Week 307Named business-affairs executives to approve
→Executive approval workflow, then pilot and production validation
Weeks 5–6
Nothing else is required
Deployment, documentation and Agent Care handover are ours.
Delivery timeline
Four phases across six weeks
The bands are drawn over the weeks each phase genuinely needs, so the fifth doubles.
Phase
W1
W2
W3
W4
W5
W6
- DiscoveryW1
- BuildW2 – W3
- EvaluateW4 – W5
- Pilot & LaunchW5 – W6
Week focus
W1Consent workflow discovery, scope mapping and the automation boundary
W2Register, casting and VFX integration and the match baseline
W3Matching workflow, confidence logic and approval controls
W4Evaluation suite, consent-artefact checks and failure-mode testing
W5Asset and contract-system integration, pilot uses and corrections
W6One production run under the business-affairs executive, then Agent Care handover
- Reading the band
- Each bar covers only the weeks its work is named in. Week 5 carries two phases because it does.
- At the end of W6
- Validation closes on live consents, and Agent Care picks up monitoring.
- Duration
- Six-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.
Next step · Media & Entertainment AI agent
Build a digital-replica consent agent around your business-affairs file.
Show us your capture sources, your consent forms and who approves a use. If any replica on your slate was built from production footage rather than a scan, the consent regime already reaches it.
- Typical delivery
- 4–6 weeks
- Deployment
- Your register and contracts
- Executive approval
- Per-use consent
- After launch
- Agent Care available
Nestack Agents · Digital replica & consent
AGT-ME-13 · Agent Care available after launch