For years, Artificial Intelligence (AI) has been sold to marketers as a productivity tool. Across Asia-Pacific (APAC), however, that description is already proving too narrow. Consumer behavior has never been more dynamic. Customer expectations shift in real time, campaigns run continuously across channels, and competitive advantage is increasingly measured in minutes rather than days. Marketing leaders are expected to respond at the same pace.
That’s changing the role AI plays inside modern marketing. Rather than simply helping marketers work faster, AI is taking on operational decisions that humans no longer have the capacity to continuously make themselves, from monitoring campaign performance and detecting anomalies to protecting ad spend around the clock. An example is how ChatGPT reaches over 900 million weekly active users and launches ads globally – AI is rapidly becoming both the operator behind campaigns and a major channel itself.
Our data confirms this shift is already well underway in APAC. The region’s mobile-first consumers, highly competitive digital economy, and rapid adoption of AI-powered marketing technologies have made it an early indicator of where marketing is heading globally. The region now accounts for nearly half (49%) of global Model Context Protocol (MCP) connections, the connective layer that allows AI agents to interact with external tools and data sources. Across the region, agents are running at 70% of weekday volume on weekends.
The weekend statistic is perhaps the clearest signal that something fundamental has changed. When marketing systems continue operating at 70% of weekday activity after teams have logged off, AI has moved beyond automation. It has become part of the operating model itself.
The question is no longer whether organizations are adopting AI. It’s whether their measurement capabilities are evolving quickly enough to support the decisions AI is already making.
Speed is the headline. Foundation is the story.
AI agents are already demonstrating enormous operational value. They can monitor campaigns continuously, surface anomalies, and react to signals far faster than any human team could. That capability will only become more important as consumer behavior becomes increasingly dynamic and marketing decisions need to be made in near real time.
Our data suggests this shift is accelerating rapidly. Token authentication, the technical marker that distinguishes automated agentic workflows from human-initiated ones, surged from 14% to 48% of all activity in just six months. That means AI is increasingly operating independently: scheduling data pulls, monitoring campaign performance, detecting anomalies and protecting ad spend around the clock, without waiting for someone to log in.
The commercial value of this is real and, in some cases, dramatic. One gaming team’s always-on agent, for example, detected a budget pacing issue at 2 AM on a Saturday that would have silently drained 40% of their weekend spend before any analyst came online to see it. That kind of intervention is only possible because the agent was continuously watching through a window that no human team could sustainably cover.
As AI agents take on greater responsibility for monitoring performance, reallocating spend and surfacing recommendations, however, the quality of those decisions becomes inseparable from the quality of the data they’re built on. The value of AI no longer depends simply on how quickly it can act, but on whether it is acting on signals organizations can trust.
That’s why the conversation about AI can no longer focus solely on speed or automation. The challenge isn’t making decisions faster. It’s making decisions you can trust at the speed modern marketing demands. When agents make decisions at machine speed across campaigns and audiences, marketers need confidence that those decisions are grounded in real, trusted data. If any uncertainty already exists within the current measurement process, AI will only amplify it. If the underlying data is flawed, automation won’t correct it; it will simply scale those flaws faster than human teams can identify and intervene.
What separates successful AI adoption from expensive automation
It’s tempting to assume that competitive advantage comes from deploying more sophisticated AI models. What we’re seeing suggests something different. The organizations extracting the greatest long-term value from AI aren’t necessarily deploying the most advanced tools – they’re building them on stronger foundations.
With 53% of MCP accounts returning on day 7 or later, there is clear evidence that sustained value is being created. The workflows driving that retention share a common trait: they invested in trusted measurement and clean data before layering AI on top.
AI may accelerate decision-making, but the quality of those decisions will always be limited by the quality of the signals feeding them.
What this means for APAC’s core industries
While the underlying challenge is the same across industries, the consequences look very different depending on what organizations are asking AI to do. Across some of the region’s largest digital sectors, one pattern is becoming clear: the more responsibility AI takes on, the more confidence in the underlying data becomes a competitive advantage.
In eCommerce and retail, the advantage is increasingly one of decision velocity. AI can join attribution data with first-party signals, identify performance shifts and recommend where to move the budget in minutes rather than hours. The shopping journey has also changed fundamentally; discovery now happens across social feeds, live commerce and AI-driven search, raising the bar on what measurement has to capture. As Jane Hou, Industry Lead for eCommerce and Retail, APAC at AppsFlyer, puts it: “In eCommerce, more of the journey now happens across social, live and AI-driven discovery. Speed isn’t the hard part anymore — it’s making sure the data keeps pace with how people actually buy before you let AI act on it. The brands that invest in that get a real edge; the ones that don’t just reach the wrong decision faster.”
For Banking, Financial Services and Insurance (BFSI) marketers, the conversation is no longer just about optimizing campaigns — it’s about accountability. As AI becomes more deeply embedded across the customer lifecycle, every automated decision needs to be explainable and auditable to internal stakeholders and regulators. Trusted measurement isn’t simply a performance metric; it’s the foundation for governance and compliance. Rather than becoming a barrier to AI adoption, strong regulatory frameworks present an opportunity for financial institutions to streamline onboarding, build customer trust, and acquire higher-quality, higher-lifetime value (LTV) customers with greater confidence. As Ribo Alam, Industry Lead for BFSI, APAC at AppsFlyer, puts it: “In financial services, AI raises the stakes. Every decision now needs to be effective for the business and explainable to regulators. Trusted measurement is the foundation for that. Done right, the compliance question and the growth question have the same answer: clean data, auditable decisions, and customers you can actually stand behind.”
Gaming has long been among the first industries to adopt emerging technology, from programmatic advertising to creative automation, and AI is no different. Every optimization maps directly to user acquisition, retention, and lifetime value; gaming offers the clearest glimpse of where every industry is heading. AI agents are already monitoring campaigns continuously, protecting budgets and responding to anomalies long before marketing teams come online. Supersonic Studios, a mobile gaming publisher running large-scale user acquisition campaigns, learned this firsthand. When raw data was fed directly into a large language model (LLM), the outputs did not match their own dashboards, creating a reliability problem that undermined the entire workflow. The fix was to pre-aggregate on the business intelligence (BI) side first, establish a clean and trusted signal, and then build the agent layer on top. Once they did, the outputs aligned. As Wilson Yen, Industry Lead for Gaming, APAC at AppsFlyer, puts it: “Gaming never sleeps, and neither do the agents we’re deploying. The ones that actually protect, spend and catch problems at 2 am are built on clean, trusted data. The sophistication of the model matters far less than the reliability of the signal it’s watching.”
What should APAC marketers do now?
The challenge described above of AI moving faster than the measurement infrastructure beneath it is not inevitable. The competitive advantage in this next phase of AI adoption will go to the teams whose foundations are strong enough to trust the decisions their agents are making, and three priorities stand out for marketers in the region right now.
- Build your measurement layer before your AI layer: audit your data signal quality before scaling agentic workflows. Automation finds your blind spots faster than you do, and knowing where your measurement gaps are before agents start acting on them is the difference between catching a problem early and locking in the wrong decisions at scale.
- Learn to measure humans and machines differently: Token-authenticated workflows and conversational usage have meaningfully different profiles that need to be analyzed separately, because mixing them produces a blended picture that obscures what is actually driving performance.
- Treat measurement as a strategic capability: Measurement is no longer simply about reporting campaign results. As AI becomes more embedded in day-to-day marketing operations, it becomes the mechanism that allows organizations to validate, govern, and ultimately trust the decisions their AI is making.
APAC has moved faster than any other region into the agentic era. The organizations that lead the next phase will be the ones with the confidence to trust the decisions their AI is making. In an environment where consumer behavior, campaigns, and AI agents are all moving faster than ever, trusted measurement is no longer a reporting exercise. It’s the foundation that allows marketers to move quickly without sacrificing confidence.
This thought leadership piece is written by Ziv Peled, Chief AI and Customer Officer at AppsFlyer
