A reference guide for agencies, brands, and production teams
AI-enabled production introduces new roles, cost structures, and contractual considerations that don't yet have consistent language across the industry. Without shared terminology, scope gets misunderstood, value gets misattributed, and protections that should be standard are treated as one-off asks.
This glossary exists to fix that. It draws on emerging industry standards being developed collaboratively across production bodies in the UK and the US, and on BearJam's real-world experience across agencies, direct brands, and international production.
We update it as the industry evolves. If a term you're looking for isn't here yet, get in touch.
The proprietary know-how a studio uses to combine multiple AI tools, techniques, and specialists into a single, coherent production system. It includes workflow design, tool orchestration, handoff standards, creative controls, QA gates, automation scripts, versioning conventions, and the judgment required to choose the right tool at the right moment, making disparate outputs behave like one finished piece of work.
Also referred to as: Creative Operating IP, Production Craft IP.
The metered technical resources used to generate, process, and deliver AI outputs: model usage (tokens and credits), GPU/CPU time, memory, storage, bandwidth, and the software runtime needed to execute generation, simulation, rendering, upscaling, compositing, and encoding.
Analogous to render farms or film stock in traditional production. Output volume scales roughly linearly with compute cost. Should appear as an explicit line item in any AI production budget.
The structured process for routing AI-generated outputs through client approval. Includes stakeholder mapping, consolidated feedback windows, revision tracking, and scope protection for out-of-process feedback. AI production generates significantly higher output volume than traditional workflows, making review management more complex and more important to scope explicitly.
Additional legal oversight required to manage contractual, regulatory, and IP risks specific to AI-enabled production. This includes reviewing AI-related clauses (ownership, licensing, training-use restrictions, confidentiality, warranties and indemnities, talent rights, and usage rights), assessing provenance for inputs and outputs, advising on disclosure language, and documenting due diligence steps such as model and tool policies, audit trails, and risk assessments.
A fee allocated to quantify and mitigate the estimated greenhouse-gas emissions associated with compute used on a project: model runs, rendering, storage, and data transfer. Carbon offset fees are typically applied via accredited carbon credits or verified climate projects intended to counterbalance a defined portion of the project's emissions footprint.
Measures and infrastructure required to meet client security requirements: access controls, secure storage, encrypted transfer, isolation, audit logging, retention and deletion policies, and vendor vetting where applicable.
Direct costs for third-party tools required for a project: model licences, creative software, plugins, render tools, and asset libraries, charged as pass-through or per a rate card. The licence tier matters: enterprise licences typically grant full commercial IP rights and restrict training-data ingestion.
The effort and systems required to document inputs, tool and model versions, permissions, and key production decisions, sufficient to support IP verification, regulatory compliance, and liability management. Increasingly required by regulated industries and large brand clients.
Creative leader responsible for concept development, brand adherence, and creative quality of AI-generated outputs. Directs AI Artists and sets the visual and narrative direction. This role steers, curates, and approves, distinct from hands-on generation. As the marginal cost of generation approaches zero, the differentiator is taste and judgment: this is where it resides.
Reference gathering, mood boards, style guides, presentation materials, and visual quality control. May also generate AI outputs, but primary value is in curation, context-setting, and ensuring visual coherence across high-volume output.
Creative practitioner responsible for generating and shaping outputs using AI tools and references: visual, audio, or motion. Work includes concepting, iterative exploration, curation, and refinement to meet the brief and brand standards, often coordinating closely with VFX and finishing teams.
Owns final image and sound integrity and delivery readiness. Ensures continuity, compositing, colour, timing, typography, codec and specification compliance, and that mixed-source outputs feel intentional and premium.
Technical practitioner responsible for building, adapting, and running AI-enabled production workflows. Work includes toolchain integration, automation, environment setup, model evaluation, and when required, fine-tuning, training, and deployment of models or components.
Senior technical owner responsible for designing, operating, and safeguarding the end-to-end AI production pipeline. Selects approaches and tools, defines integration and handoff standards, sets QA and reliability gates, manages technical risk, and resolves escalations to ensure outputs remain coherent, brand-safe, and deliverable across the full toolchain.
Hybrid practitioner, both creative and technical, who prototypes, tests feasibility, and defines the 'how' early on. Translates creative ambition into practical workflows, constraints, and options. Produces proof-of-approach prototypes, de-risks unknowns, and shapes the brief into something executable.
Owns schedule, scope, budget, and stakeholder alignment for AI-enabled production. Manages iterations, review loops, approvals, vendor coordination, and delivery packaging. Keeps creative, technical, and finishing teams moving in sync.
This glossary is developed in alignment with emerging standards from the APA (UK) and AICP (US). It will be updated as the industry's shared language evolves.
See our AI production work at bearjam.co.uk/ai-video-production. Questions or additions: hello@bearjam.co.uk