PT Cerita Inovasi Kopi Indonesia (CIK) is an Indonesian coffee company running a growing, multi-outlet retail operation. Its business is already digitized on an ERP platform covering procurement, inventory, sales, finance, CRM, reporting, and internal approval processes. As the company expanded, its core systems were not the constraint — the bottleneck moved to the work still done by hand: producing visual content for campaigns and compiling operational reports from data the systems already held.

● Content Production Bottleneck: Every campaign and promotion required different visual assets for every outlet. Production depended on designers and took several days to yield a single draft, while demand kept rising as the outlet count grew.
● Per-Asset Costs: Each asset briefed to a designer or external vendor carried a per-asset cost, and many campaign ideas launched late or were cancelled outright because the visual material was not ready.
● Time-Consuming Reporting: Data had to be gathered from the ERP, individual outlet reports, and various spreadsheets before analysis could begin — roughly 45–60 minutes for every routine report.
● Data Security Requirement: All data must stay inside the company environment and must never be used to train public AI models.

CIK needed to produce campaign images and videos quickly and at low marginal cost, cut manual effort in routine reporting, and integrate with its existing ERP without any re-platforming. Data security was decisive: operational, customer, and financial data had to remain private and never be used to train public AI models.

QwenWork, Alibaba Cloud's AI Agent Workforce, met these conditions:
● Private Environment: Runs entirely within CIK's environment with encrypted data processing.
● Data Residency: Hosted in the Alibaba Cloud Jakarta Region for low latency and compliance.
● Advanced AI: Powered by Alibaba's Qwen3.7-Max model.
● Ease of Use: Natural-language conversational interface lets non-technical staff use the agents directly.

CIK leverages QwenWork to deploy two AI agents for marketing content creation and operational reporting.

● Marketing Content Assistant: Generates promotional images and videos from a brief — social media campaign material, menu launches, seasonal promotions, and outlet-specific visuals — drawing on CIK's brand guidelines and product catalogue so output stays consistent with brand identity. Every asset passes human review before publication.
● Reporting Assistant: Drafts operational report summaries through managed read-only access to ERP and outlet system data, accelerating report preparation without altering the ERP, which remains the system of record.

● Reporting Efficiency: 30–50% target efficiency gain. Compiling one report previously meant manual work across the ERP, outlet reports, and spreadsheets, taking 45–60 minutes. The Reporting Assistant drafts summaries in minutes, reducing the task to a few minutes of review.
● Content Production Efficiency: 30–50% target efficiency gain. Assets previously briefed to a designer or vendor took days per batch and carried a per-asset cost; they are now generated in-house from a prompt.
● ERP Integration: The AI Agents work on top of the ERP already in place — no re-platforming required.

CIK plans to leverage Qoder, Alibaba Cloud's AI coding assistant, to accelerate internal application development and system integration. The two agents already in production will expand across more campaigns and outlets as the network grows.