As artificial intelligence becomes more deeply embedded in marketing, single-function agents are increasingly proving inadequate for real-world use cases, which tend to be too complex and interconnected for narrowly scoped tools.

That limitation was a recurring theme at the 2026 iFlytek Global AI-Marketing Ecosystem Conference, held on January 15 in Shenzhen. At the event, iFlytek’s AI marketing unit, AIMarX, officially unveiled its SuperAgent application framework, alongside an agentic marketing platform built on top of it.

Data fragmentation remains the central bottleneck preventing AI from scaling in real-world marketing environments, said Ronen Mense, Asia Pacific president and managing director at AppsFlyer. He described the issue as a structural challenge facing the industry rather than a purely technical one.

As marketing touchpoints continue to multiply, even large brands struggle to integrate data collected across disparate platforms in a meaningful way. The result is that many AI applications remain confined to isolated, single-point optimizations rather than end-to-end decision-making. According to Mense, marketing teams are more exposed than most other business functions to weak data foundations. Data must comply with privacy regulations while simultaneously supporting real-time decisions, he said, a dual constraint that makes it difficult for companies to build trusted and unified data infrastructure.

These pressures are amplified in global marketing operations. “The complexity of global marketing is rising exponentially due to diverse market hierarchies, stark cultural differences, and highly dispersed touchpoints,” said Wang Yunxia, vice president of iFlytek’s AI marketing business group and general manager of its marketing solutions division. “These are problems that can no longer be solved by simply adding more human labor.”

“AI is evolving from a function into foundational infrastructure,” said Yu Jidong, iFlytek’s senior vice president and president of its consumer business group. This transition, he added, calls for a more pragmatic approach to technology, arguing that only solutions tied directly to real business pain points can drive meaningful change across the industry.

Discussions throughout the conference pointed to a growing consensus that isolated technological advances are no longer sufficient to address systemic challenges. What is needed, participants suggested, is reconstruction at the ecosystem level.

That process begins with how teams work together. Traditional marketing organizations are typically siloed, with data analysis, creative development, and media buying operating largely in parallel. Intelligent marketing, by contrast, requires cross-functional and tightly integrated workflows. “Optimizers can quickly pinpoint strategy issues, while creative teams can immediately see which assets are performing well,” Wang said. “This unified perspective is fundamentally reshaping traditional collaboration models.”

A deeper shift is also underway in how marketing value is measured. According to 36Kr, 62% of users no longer click on traditional links after receiving AI-generated answers. If accurate, this trend reduces the usefulness of metrics such as clickthrough and conversion rates, which may no longer fully reflect marketing effectiveness. The implication is that the industry may need new measurement frameworks capable of capturing longer-term trust and value-based relationships between brands and consumers.

Li Ping, vice president at iFlytek and president of its AI marketing business group, expressed confidence in the company’s positioning. He said SuperAgent is intended to help marketing relationships evolve from a traditional B2C model to a B2A2C (brand-to-agent-to-consumer) structure, linking brands to consumers through AI agents. Under this framework, AI is positioned not only as a productivity tool, but as an intermediary that reshapes how brands and consumers interact.

KrASIA Connection features translated and adapted content that was originally published by 36Kr. This article was written by Ou Xue for 36Kr.