To advance the national "AI+" action plan, strengthen scenario-driven development, and accelerate intelligent upgrades in specialized services, the Jing'an District has established the "Jing'an SaS Academy." This initiative aims to cultivate the new Service as Software (SaS) business model. The academy serves as a high-quality platform for AI application deployment and industrial resource connectivity, focusing on five key professional service sectors: consulting, human resources, advertising, legal, and finance and taxation. Through themed sharing sessions, case analyses, live exchanges, and supply-demand matchmaking, the academy organizes small-scale, high-frequency, and topic-specific events to facilitate the deep, two-way integration of specialized service intelligence and intelligent service specialization.

On September 17th, the inaugural session, titled "AI-Empowered Marketing for Mutual Growth," was held at the Jingtou Center. This marketing-focused event was hosted by the District Science, Technology and Economy Commission and the District AI Task Force, with support from the District Commerce Commission, the Shanghai Industrial Internet Association, and the District Foreign Investment Enterprise Association.

The event took the form of a small roundtable salon, inviting two industry-leading companies to serve as industrial mentors. They engaged in face-to-face dialogue, addressing questions and exploring resources with representatives from eight enterprises in the marketing services and fast-moving consumer goods retail sectors. The goal was to help businesses unlock new AI-driven marketing strategies and identify new growth drivers.

In the opening remarks, the District Science, Technology and Economy Commission highlighted that specialized services represent Jing'an's distinct advantageous industry and a key arena for AI scenario implementation. Promoting the intelligent upgrade of professional services is a crucial step for Jing'an to cultivate new business forms, strengthen new growth drivers, and foster high-quality industrial development. The establishment of the Jing'an SaS Academy aims to create a professional, routine, and precise platform for AI supply-demand matching, accurately connecting technology providers with enterprise needs to address common transformation challenges such as "not knowing how," "fearing," and "underperforming" in digital adoption. The Commission encouraged attending companies to translate the insights from the discussions into actionable transformation steps, moving from listening to concepts toward practical implementation, thereby enabling more AI applications to take root and flourish within Jing'an's business community.

During the themed sharing segment, JST GROUP presented on "From 'Understanding Users' to 'Sustained Growth' – An AI-Driven Full-Chain Reconfiguration of Marketing and Sales." They systematically deconstructed the pathways through which AI reshapes brand growth logic. The presentation proposed evolving AI's role from a "tool" for completing single-point tasks and an "assistant" for coordinating people and processes, to a "digital employee" that truly takes on business responsibilities. By building an enterprise-level AI operating system based on four engines—cognition, execution, evolution, and iteration—they aim to push AI beyond providing better answers and toward successfully executing tasks. The session also showcased practical results from multi-agent collaborative services, AI sales coaching, and GEO intelligent assistant implementations.

The presentation by Jushuitan, focusing on "AI Empowering E-commerce: Jushuitan's AI Implementation Practices and Exploration," identified a common industry pain point: "AI fails to land, not because it's not intelligent enough, but because it isn't integrated into real business operations." Jushuitan's solution employs an "ERP-native AI" approach, where an AI marketing agent unifies access to advertising, order, product, inventory, and gross profit data. By overlaying business operational rules, the AI gains a complete understanding of the business context, enabling it to move from mere analysis to decision-making. Using "Shuibao" as a smart access point, the focus is on "understanding the business, being trustworthy, and being affordable," making AI a practical tool for merchants' daily operations.

The Q&A and discussion session was lively, with participating companies raising targeted questions about real challenges and operational difficulties encountered during their intelligent transformation journeys.

Question: How should the effectiveness of GEO be validated, and why are results inconsistent?

Answer: GEO operates differently from traditional SEO; as AI models continually evolve, rankings naturally fluctuate. GEO is not a one-time deliverable but a long-term, ongoing effort. Validation should focus on long-term trends rather than point-in-time checks, and performance fluctuations can be stabilized through continuous optimization.

Question: With high labor costs and repetitive queries in customer service, can AI deliver tangible cost reduction and efficiency gains?

Answer: Intelligent customer service is currently one of the most mature and clearly defined areas for AI implementation. It provides 24/7 response capability, freeing human agents from repetitive inquiries. It represents an excellent "first step" for enterprises beginning their AI journey.

Question: What is the underlying logic of AI-driven marketing, and how should brand content be provided for AI training?

Answer: Companies should establish brand content standards and codify them as knowledge assets. This means consolidating confirmed brand tone, style, and content direction into a resource library to continuously "feed" and calibrate the AI through iterative platform layers. The core principle is to teach AI the "company's specific standards" rather than relying solely on a generic model's default output.

Question: For traditional trading companies without an AI foundation, where should they start?

Answer: The recommendation is to first conduct a thorough review of departments to identify scenarios, starting with highly mechanical and repetitive tasks such as document organization or approval workflows. Begin with small, focused initiatives to achieve visible success, then gradually expand from there.

Question: With many AI Agent products available, how should companies evaluate them, and how does this differ from traditional software evaluation?

Answer: AI projects are processes of continuous evolution and learning. Agent evaluation should be results-driven, paying for actual output ("pay-for-results" model). The focus should be on whether the agent can consistently deliver positive outcomes on specific business challenges, rather than assessing a one-time delivery.