Define communication goals and the applicable boundaries of AIGC
Before launching any AIGC commercial project, the primary task is to clarify what generative imagery can and cannot solve. AIGC excels at handling abstract concepts, surreal visuals, rapidly iterating emotional atmospheres, and low-cost proof of concepts. It is not suitable for hard-sell shots requiring precise display of physical product details, complex mechanical structures, or strict adherence to brand VI guidelines. Brands must assess whether their core selling points rely on high-fidelity physical representation. If the product's appearance is the key decision factor, pure AI generation often struggles to guarantee pixel-level accuracy; in such cases, it should be positioned as supplementary material while avoiding being the main subject. This judgment directly dictates subsequent budget allocation and technical route selection, preventing the forced application of unsuitable technologies to mismatched marketing goals.
Build a brand brief that includes technical constraints
Traditional creative briefs focus only on narrative and aesthetics, but in AIGC projects, sections on technical parameters and compliance must be added. Brands must explicitly specify aspect ratio, minimum resolution, frame rate requirements, and final distribution platforms in the document. More importantly, prohibited elements must be listed, such as specific competitor logos, sensitive social issues, or color tendencies that do not fit the brand tone. Additionally, the standard for "consistency" must be defined, meaning the similarity tolerance for characters, scenes, or products across different shots. This document is not only a guide for the creative team but also the legal and technical basis for post-production review; missing these constraints will result in generated outputs failing internal legal review or brand standard validation.
Sample validation, style locking, and asset testing.
Before entering formal production, the style sample validation phase must be completed. The goal of this phase is not to produce the final cut, but to test the AI model's ability to reproduce specific brand elements. The team should select the three most challenging shots for generation testing, focusing on lighting logic, material texture, and text generation accuracy. If facial distortion, blurred product edges, or illogical backgrounds are found, immediately adjust the prompt engineering or switch workflows. The deliverable for this phase is a set of confirmed reference frames to serve as a baseline for subsequent large-scale generation. Skipping this step and going straight to mass production will highly likely result in a fragmented overall visual style, causing irreversible time waste and cost overruns.
Hybrid workflow integrating live-action and CGI.
Most successful AIGC commercials are not entirely AI-generated, but use a hybrid workflow. Pre-production shooting must leave enough compositing space for post-production, meaning lighting setups must match the virtual light source directions generated by AI, and actors' performances must accommodate interacting with non-existent objects. The production team should annotate during the storyboard phase which parts use live-action and which parts use AI extension or replacement. For example, use live-action to capture real human performances and interactions, then use AI to generate complex background environments or VFX elements. This division of labor requires the director of photography and the AI artist to collaborate closely to ensure consistent perspective, motion trajectories, and color science between the two, otherwise obvious inconsistencies will be exposed during post-production compositing.
Standardized cleanup and repair in post-production.
Raw AIGC-generated assets typically suffer from flickering, artifacts, or illogical details. The post-production team's primary task is not editing, but repair. This includes removing timeline jitter, correcting finger counts, cleaning background noise, and unifying color tones. For projects involving human faces, pay special attention to the naturalness of micro-expressions, and use traditional CGI techniques for repainting if necessary. Additionally, subtitles and voiceovers must be re-synced, as AI videos often have audio-visual sync issues. The investment in this phase is often underestimated, but it is critical to the professionalism of the final cut. Delivering AI assets without fine-tuned repair directly will be perceived by audiences as low-quality content, severely damaging the brand image.
Copyright clearance and compliance review.
Before using any generative tool, you must confirm whether its license agreement permits commercial use. Many free or subscription-based platforms have strict restrictions on the copyright ownership of training data, and may even retain control over the generated content. Brands must collect source proof for all generated assets to ensure no risk of infringing on third-party intellectual property rights. For projects involving real people, if deepfake technology is used, written authorization from the individuals must be obtained. Legal counsel should be involved in this phase to create a written record. Ignoring copyright risks may result in the video being taken down after release, or even facing lawsuits, with consequences far exceeding the production cost itself.
Multi-version adaptation and delivery manifest management.
Deliverables for AIGC projects are far more complex than traditional video. In addition to the final master file, adapted versions for different platforms must be provided, such as vertical short videos, horizontal long videos, and static poster assets. Each version must be individually checked for composition cropping and text readability. The delivery manifest should include source project files, uncompressed intermediate formats, color grading LUTs, and font packages. If the project involves team collaboration, permissions and modification history at each level must also be noted. Clear delivery management reduces communication costs and ensures the operations team can smoothly access the assets. A lack of standardized delivery criteria leads to file chaos, increasing the cost and time of secondary production.
Acceptance criteria and risk control mechanisms
Acceptance should not rely solely on subjective aesthetics; a quantified checklist should be established. First, verify technical metrics, such as whether resolution, bitrate, and audio levels meet publishing requirements. Next, check content consistency to ensure the brand logo and product features are accurate across shots. Then, assess compliance to confirm there is no sensitive content or copyright flaws. Finally, conduct multi-platform playback testing to simulate effects across different network environments and device screens. If major defects are found, return to the corresponding stage for rectification. Establishing such a complete process mechanism can effectively reduce delivery risks. For content that does not meet standards, there should be a clear rejection and rework process to avoid launching with issues.
When AIGC solutions are not applicable
Although AIGC has huge potential, it is not applicable in certain scenarios. When a brand emphasizes extreme craftsmanship, authentic physical interaction, or highly customized emotional connections, the generic aesthetics generated by AI may appear thin and soulless. Additionally, if a project timeline is extremely short and requires an immediate response to trends, but the team has not yet established a mature AI workflow, forcing its adoption may lead to quality loss. Similarly, if the audience is skeptical of new technologies or the brand positioning leans traditional and conservative, overusing AI may trigger negative public opinion. In these cases, returning to traditional live-action or mixing in minimal AI effects may be the safer choice. Rationally evaluating the fit between technology and the brand is more important than blindly chasing trends.
Next step recommendations
It is recommended that brands start with small-scale concept short films to accumulate internal experience and data. Form a cross-functional team including creative, technical, legal, and operations personnel to jointly develop suitable working standards. Before official project approval, conduct a complete full-process simulation to identify potential bottlenecks. The services publicly available on the ONCE official website cover corporate promotional videos, brand promotional videos, TVC commercials, product videos, overseas marketing videos, social media short videos, and AIGC video production, which can serve as a reference option for professional support. Maintaining sensitivity to technological iteration while holding firm to core brand values is the only way to move forward steadily amid change.
If you are preparing an AIGC commercial project, you can first organize the brief, reference visuals, product or corporate materials, delivery platforms, and copyright scope, then review theAIGC video services pageto ground communication from abstract preferences to executable production boundaries.
