I. Two Core Types of AI Ad Videos
According to the 2026 China AI Marketing Market White Paper published by iResearch Consulting, product sales-type ad videos account for 58%, while brand promotion-type accounts for 42%. The production logic and cost structure of the two types differ significantly:
Product Sales-Type Videos:
The core objective is to drive conversions, typically placed on e-commerce detail pages or in-feed ads. The average conversion rate is 3.8%, requiring the core product selling point to be showcased in the first 3 seconds. The production cycle is short (3–5 days), with a per-unit budget of 1,000–3,000 RMB.
Brand Promotion-Type Videos:
The core objective is to enhance brand awareness and favorability, typically used for social media and brand website placements. The average video completion rate is 12 percentage points higher than product-type (42% vs 30%). The production cycle is longer (7–15 days), with a budget range of 5,000–20,000 RMB.
II. Five-Step Workflow Explained
Step 1: Goal Positioning and Audience Analysis (Day 1)
Define the core KPIs for the video—are you pursuing click-through rate, conversion rate, or brand search index growth? Choose the visual style based on the target audience profile: younger demographics prefer fast-paced editing and trending music, while business audiences lean toward subdued color tones and professional narration.
Step 2: Script Planning and Storyboard Design (Days 2–3)
Use a “three-act structure” to write the script: opening hook (create suspense or visual impact within 5 seconds) → core message delivery (product showcase / brand story, occupying 70% of the video duration) → call to action (direct viewers to click a link or search for the brand name).
For storyboard design, it is recommended to switch the visual rhythm every 3–4 seconds to avoid viewer fatigue. Use ProcessOn or Figma to create simple storyboard sketches, annotating each shot's content and duration.
Step 3: AI Image Generation (Days 4–6)
Static Image Generation Tool Recommendations:
Nano Banana Pro is suitable for realistic product scene rendering and supports 4K resolution output; Seedream 5.0 excels at abstract backgrounds and emotional atmosphere creation, making it particularly suitable for beauty and luxury advertising.
GPT Image 2 (called via the ComfyUI-relayapi node) leads in text rendering capability, able to generate brand slogans or promotional information directly within the image without the need for post-production editing in Photoshop.
Video Animation Conversion Tools:
The LTX-2.3 model supports precise camera movement control (pan/zoom/orbit), making it suitable for 360-degree product showcase shots; Seedance 2.0 has clear advantages in multi-character scenes and character motion generation, with an average lip-sync error of ≤3 frames.
Pro tip: First generate a high-quality static reference image → import it into the video model to add motion trajectories → adjust the “motion strength” parameter to between 0.6–0.8 (values too high will cause image distortion).
Step 4: Voiceover and Sound Effects Synthesis (Day 7)
TTS Voice Synthesis:
Qwen-TTS offers 20+ voice options; “Energetic Female – Xiaoya” is recommended for FMCG ads, while “Steady Male – Weijie” suits tech products; Doubao TTS has the richest dialect support (Cantonese/Sichuan dialect/Northeastern dialect and 15 others), making it ideal for regional marketing campaigns.
Recommended speech rates: for information-dense product videos, use 1.2x speed to ensure complete message delivery; for emotional story-type ads, use 0.9x slow speed to create an immersive experience.
Background Music and Sound Effects:
Suno AI can generate custom soundtracks—input keywords like “upbeat, corporate, electronic” to produce a 30-second copyright-free instrumental track. Ambient sound effects (footsteps/door closing/product click sounds) are recommended to be downloaded from the free Freesound.org library.
Step 5: Post-Editing and Multi-Platform Adaptation (Days 8–9)
Use Jianying Professional Edition or DaVinci Resolve to complete the final composite. Focus on checking the following quality indicators:
Lip-sync error ≤3 frames (approximately 100 milliseconds); color consistency score ≥85 (using the Adobe Color Grading tool to check); audio waveform peak levels controlled between -6dB and -3dB to avoid clipping.
Output formats: landscape version (1920×1080, MP4 encoding, 8Mbps bitrate), portrait version (1080×1920), and square version (1080×1080). File size should be kept under 50MB to ensure smooth loading on platforms like Weibo/WeChat.
III. Cost-Efficiency Comparison
Total cost of producing a 60-second standard ad video using the AIGC approach:
Creative planning fee: approximately 500–1,000 RMB + AI tool API call fees: 300–800 RMB + manual review and refinement fees: 800–1,500 RMB = approximately 1,600–3,300 RMB total. Production cycle compressed to 7–9 days.
A traditional live-action shoot of equivalent quality costs 20,000–50,000 RMB (including venue rental/actor fees/lighting equipment/post-production effects). The AIGC approach can reduce budget expenditure by 85%.
IV. Common Technical Troubleshooting
Issue 1: AI-generated images show limb deformation or abnormal fingers
Solution: Add negative prompts like “bad hands, bad anatomy” in the prompt; or use ControlNet's OpenPose module to lock the character's skeletal structure.
Issue 2: Inconsistent character appearance across multiple shots
Solution: Consistently use the same character reference image as the input baseline for all scenes, and enable the model's “--cref” (Character Reference) function to maintain 90%+ facial consistency.
Issue 3: Audio-video desync after video export
Solution: Check whether the audio track's starting frame on the Jianying timeline is aligned with the visual keyframes; if the issue persists, try switching the output encoding format from HEVC to H.264.
V. Delivery Acceptance Checklist
Resolution no lower than 1080P; duration deviation within ±3 seconds; lip-sync error ≤3 frames; color space sRGB or Rec.709; file size no larger than 50MB (Weibo platform limit).
Visit the AIGC SDM supply-demand matching platform AI Ad Video Custom Services to obtain professional-grade AIGC ad production solutions.
