What challenges did Akool face before AI implementation?

Existing B2B AI video tools lacked stability and usability for real-world business use. Moving from offline generation to real-time interaction required low latency, efficient large-scale processing, consistent quality and realism, and elimination of flickering and product deformation while meeting enterprise brand and compliance standards.

What was the impact of the AI solution?

The article says the approach improved stability in complex scenarios and reduced character flickering in internal tests. It says the platform served millions of users across 180+ countries, generated more than 300 million creative assets, translated video into over 180 languages and dialects in minutes, and cut localization costs and time. It also cites 40-60% cost reduction versus pure pay-as-you-go models through flexible deployment options.

Source

AlibabaEvidence: Medium50/100

Akool builds production-grade B2B AI video generation on Alibaba Cloud with Qwen-VL and Model Studio

Use case typeContent generationUpdated Jul 19, 2026

Akool is a Silicon Valley-based B2B AI video generation company serving global enterprises. The company built a production-grade AI video architecture on Alibaba Cloud using Alibaba Cloud Model Studio, Qwen-VL, and Wan2.6 to improve stability, continuity, multimodal understanding, and compliance for real-time and high-fidelity video generation.

Organization: Akool
Industry: Consumer & Food
Location: United States
Published: July 2026

Reported outcomes

300 million

creative assets generatedOther quantified impact

40-60%cost reduction versus pay-as-you-go

Strategic outcomes

Scale & capacityReached global production scaleOther strategic outcomeEnabled rapid video translation across languagesCost efficiencyReduced localization costs and timeOther strategic outcomeImproved stability and reduced flickeringOther strategic outcomeReduced character flickering in internal tests

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Cost reduction versus pay-as-you-go: 40-60% decrease

Alibaba Cloud CustomersJul 19, 2026Customer storyExplicit claimMedium evidence strength

40-60% cost reduction compared to pure pay-as-you-go models through flexible deployment options like yearly subscriptions and reserved instances

Normalized claim

Creative assets generated: 300 million increase

Alibaba Cloud CustomersJul 19, 2026Customer storyExplicit claimMedium evidence strength

Generated more than 300 million creative assets

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: Akool
Provider: Alibaba
Maturity: Unknown
Linked source: Alibaba Cloud Customers

No explicit deployment-stage evidence found.

Customer identity supportedSource describes one deploymentMaturity evidence evaluated

Primary read

Use case focus

Showing 3 of 3

1Content generation
2Multimodal analytics
3Compliance automation

Existing B2B AI video tools lacked stability and usability for real-world business use.
Moving from offline generation to real-time interaction required low latency, efficient large-scale processing, consistent quality and realism, and elimination of flickering and product deformation while meeting enterprise brand and compliance standards.

Akool built a robust AI video generation architecture on Alibaba Cloud integrating multiple advanced models.
It used a generation-plus-understanding closed-loop system for continuity, multimodal understanding for object recognition and contextual interpretation, high-resolution image generation with style control and multilingual image-text fusion, and Qwen-VL for tag extraction, scene recognition, and compliance auditing.

Technologies

Alibaba Cloud Model Studio Qwen-VL Wan2.6

The article says the approach improved stability in complex scenarios and reduced character flickering in internal tests.
It says the platform served millions of users across 180+ countries, generated more than 300 million creative assets, translated video into over 180 languages and dialects in minutes, and cut localization costs and time.
It also cites 40-60% cost reduction versus pure pay-as-you-go models through flexible deployment options.

Architecture

Akool built a production-grade AI video generation system on Alibaba Cloud using Alibaba Cloud Model Studio, Wan2.6, and Qwen-VL. The design combines multi-scene generation, multimodal understanding, and a generation-plus-understanding closed loop to preserve character and product continuity, improve realism, and support compliance auditing.

Sources & evidence1

Evidence: Medium50/100Evidence strength