Artificial intelligence continues to transform the digital marketing landscape, and AI agents are quickly becoming the next big conversation. But as businesses race to adopt increasingly sophisticated AI technologies, an important question remains: Is the data behind these technologies ready to support them?
This year, Ayudante APAC participated in Digital Marketing Asia (DMA) Singapore 2026, a two-day event bringing together industry leaders, innovators, and marketing professionals from across the region.
While much of the excitement surrounding AI focuses on automation and efficiency, Naohiro's presentation highlighted something less visible, yet equally important: the reliability of marketing measurement.
His central message was simple: Before businesses can unlock the full potential of AI, they must first ensure that the data powering it can be trusted.
We also took the opportunity to showcase MATSUBA, Ayudante's newly launched proprietary first-party measurement service, designed to help businesses regain control of their marketing data. More on that below.
AI has become an integral part of digital marketing, from predicting customer behaviour and identifying high-value audiences to optimising campaigns and personalising experiences.
But behind every intelligent prediction lies something fundamental: data.
Naohiro revisited the familiar marketing funnel — awareness, consideration, and conversion — to explain how AI relies on two critical measurement signals:
Together, these signals help AI systems understand customer journeys and make predictions that drive marketing performance.
Naohiro walked the audience through a simple chain:
Reliable Signals → Trusted Data → Better AI Decisions → Business Impact
However, what happens when those signals are incomplete?
Even the most advanced AI models cannot accurately interpret customer behaviour that was never properly measured.
"No reliable measurement. No reliable prediction. No reliable AI impact."
As Naohiro emphasised, AI can amplify the signals it receives, but it cannot automatically compensate for gaps in the data foundation.
One of the most striking findings shared during the presentation came from Ayudante's own client investigations.
For some advertisers:
These findings highlight a growing challenge for digital marketers: the problem is not necessarily a lack of data, but the difficulty of collecting it reliably.
When measurement signals are missing, Cost Per Acquisition (CPA) and Return on Ad Spend (ROAS) figures may not fully reflect actual campaign performance. More importantly, AI-powered optimisation systems are left working with an incomplete picture.
As Naohiro explained:
"The biggest problem is that neither AI nor humans can see the data that has been lost."
And when missing data is invisible, businesses may not even realise how much their measurement gaps are affecting their marketing decisions.
One of the biggest contributors to signal loss is the tightening of cookie and tracking restrictions.
Naohiro highlighted three major challenges:
These challenges are particularly relevant across Asia, where iOS represents a significant share of mobile users in markets such as Singapore, Malaysia, Thailand, Vietnam, and Indonesia.
When conversion signals disappear, marketers face greater difficulty accurately measuring CPA, ROAS, and overall campaign effectiveness.
More importantly, incomplete data can affect the AI-powered bidding, targeting, and optimisation systems that rely on these signals.
The good news is that businesses have several ways to address measurement gaps.
Naohiro outlined five approaches:
However, these solutions are not simply switches that marketers can turn on.
Reliable measurement requires coordination across websites, browsers, tags, first-party servers, cloud infrastructure, advertising platforms, CRM systems, and data warehouses.
With AI agents becoming increasingly capable of executing complex tasks, Naohiro explored whether they could eventually manage the entire marketing measurement process.
The potential is significant.
AI agents can support measurement operations by:
However, implementing reliable measurement involves more than technical execution.
Connecting systems is complex. Deciding how those systems should operate responsibly is an entirely different challenge.
This led to one of the presentation's most important distinctions:
AI agents may automate technical implementation, but businesses still need human judgment to determine which outcomes matter, what data should be collected, and how privacy and security risks should be managed.
Importantly, greater AI autonomy does not mean unlimited access. As AI agents take on more responsibilities, organisations must establish clear permissions, approval processes, and accountability to ensure that automation operates within appropriate boundaries.
AI agents bring speed and consistency, while humans remain responsible for purpose, judgment, and governance.
As Naohiro explained, the future is not about choosing between humans and AI. It is about combining the strengths of both.
"AI agents enable. Humans govern."
Another key takeaway from the session was that a resilient data foundation cannot be built through technology alone.
Naohiro introduced three essential pillars that organisations must consider:
These responsibilities extend beyond the marketing team. Leadership, legal, privacy, IT, data, and MarTech teams must work together to establish a shared measurement framework.
This is especially important across Southeast Asia, where privacy requirements and the treatment of cookies and online identifiers differ between markets.
AI agents may accelerate implementation, but human oversight remains essential to ensure that measurement remains accurate, responsible, and aligned with business objectives.
Ultimately, building a resilient measurement foundation is as much an organisational challenge as it is a technical one.
Building on the various solutions to address measurement signal loss discussed earlier, Naohiro also took the opportunity to spotlight MATSUBA (松葉, Japanese for pine needle), Ayudante's newly launched proprietary first-party measurement service, designed to help advertisers regain control of their marketing data.
Described as a Data Ownership Engine, MATSUBA is designed to help advertisers take greater ownership and control of their marketing measurement data.
Unlike conventional browser-based tracking approaches, MATSUBA does more than deliver platform tags differently. It creates advertiser-controlled measurement records using first-party server infrastructure.
Key capabilities include:
As Naohiro emphasised:
"MATSUBA is not an AI agent. It is an expert-built first-party measurement layer for the AI era."
Its objective is not simply to recover missing conversions, but to create a more dependable measurement foundation that businesses can own, control, and use to support AI-powered optimisation.
During the presentation, Naohiro shared a case study demonstrating how MATSUBA recovered conversions that conventional platform tracking had missed.
Across three months, the case study recorded additional measured conversions of:
Note: The comparison includes platform tag tracking with Enhanced Conversions and other supported conversion types, using the last-click attribution model in Google Ads. Results may vary depending on implementation and advertising environment.
These results highlight the potential impact of recovering previously missing conversion signals.
More complete measurement does not automatically guarantee stronger business performance, but it provides advertising algorithms and marketers with a more reliable foundation for making decisions.
Importantly, Naohiro also highlighted that there is no single solution that addresses every measurement challenge.
While Consent Mode, Enhanced Conversions, Google Tag Gateway, and server-side GTM address different aspects of signal loss, MATSUBA places particular emphasis on advertiser-owned measurement.
The right approach ultimately depends on each organisation's infrastructure, privacy requirements, and business objectives.
As AI continues to evolve, it is easy to focus on what the technology can do next.
But perhaps the more important question is whether our existing data foundation is ready for it.
Naohiro's presentation served as a reminder that successful AI adoption requires more than sophisticated algorithms or automated workflows. It requires reliable measurement, clear accountability, and thoughtful human governance.
Building a resilient measurement foundation is also not a one-time project. It requires continuous monitoring, improvement, and collaboration between AI agents and human decision-makers.
As measurement challenges evolve, this ongoing human–AI partnership will become increasingly important in maintaining reliable data and supporting stronger AI outcomes.
His final message captured this perfectly:
"AI investment is visible. Missing data is not."
What isn't measured never shows up on a dashboard, making it easy to overlook.
That is precisely why investing in measurement matters just as much as investing in AI.
Last year in Singapore, our message was that prediction and protection both start with measurement.
This year, we took that conversation a step further. As AI agents take on more of the hands-on work, humans must remain responsible for defining what good measurement looks like, while advertisers need greater ownership of their data.
A big thank you to everyone who attended Naohiro's keynote, stopped by our booth, and expressed interest in MATSUBA. We enjoyed connecting with fellow marketers and industry professionals, and hope to see you again at DMA next year!
As the industry moves further into the age of AI agents, Ayudante APAC looks forward to helping more businesses bridge the gap between technological innovation and reliable, responsible measurement.
Interested in strengthening your marketing measurement foundation or learning more about MATSUBA? Get in touch with Ayudante APAC to explore how we can help.
