A featured contribution from Leadership Perspectives: a curated forum reserved for leaders nominated by our subscribers and vetted by our MarTech Outlook Advisory Board.

foodpanda
Product-Led Growth in today's AI world


ZeMing Chan
ZeMing Chan is the Head of Growth Innovation & Product Marketing at foodpanda / Delivery Hero Group. With over a decade of strategy and product experience across food delivery, e-commerce, and travel sectors, he specialises in building customer engagement programmes and driving technology-led transformation.
Capabilities Marketing Leaders Need to Drive Growth in an AI-Enabled Landscape
With the exponentially faster pace of change in this AI age, leaders need to build a robust innovation ecosystem anchored by the right culture of rapid experimentation to stand out from competition.
In this context, product-led growth (PLG) has emerged as a core competency for all marketing leaders. While brand and performance marketing are still core parts of the CMO’s toolkit, product innovation and rigorous funnel optimisation has become increasingly vital as the most cost-effective and direct drivers of organic growth.
Here in foodpanda, we invest heavily in building best-in-class tooling for marketers - from complex recommendation systems which anticipate what our customers need (and when they need it) to AB-testing platforms and automated dashboards that make continuous improvement ingrained in our processes. We have found that when these tools are built correctly, marketers adopt the technology readily, productivity increases and business impact quickly follows. For instance, when we automated AB-testing of different vouchers in our CRM flows, we achieved double-digit growth in incentive ROI across all countries within a single quarter.
Best-in-class marketing in the AI era is no longer just about clever campaigns or optimising ROAS it is about systematically engineering the growth flywheel to operate continuously, turning everyday marketing decisions into compounding growth loops.
Balancing Experimentation Speed with Brand Consistency and Customer Experience
The key is maintaining a strict "human-in-the-loop" framework. Because generative AI makes it incredibly easy to develop and deploy countless variants, there is a dangerous temptation to test everything simply because we can. However, every customer interaction is a valuable opportunity that should not be wasted on mediocre tests. We need to focus on generating sufficiently different features with high-quality differentiation before a test ever reaches the customer.
"Best-in-class marketing in the AI era involves systematically engineering the growth flywheel to operate continuously, turning everyday marketing decisions into compounding growth loops."
To balance this quality with speed, we plan and review very carefully before creating new experiences, while designing processes that increase speed of execution. This means taking an experiment-first and data-first approach while embedding brand marketers, CX specialists, legal/compliance and GTM teams into the sign-off process for faster decision making.
In my teams, we create experimentation taskforces with the right representatives from brand, CX, data, product and growth that come together weekly to plan, review and sign off on new product features & marketing campaigns. When you establish the right operational guardrails to protect the brand and customer experience, you empower the organization to iterate for speed safely, ultimately building a much faster learning organization.
How Industry Pioneers Are Changing Marketing Technology
Marketing technology is changing both for the marketer and the customer. For the marketer, agentic workflows and increased automation now enable savvy marketers to be 2X more effective. We leverage AI for creativity and experimentation but make sure we retain control of the customer experience - rather than having growth marketers manually analyze every data point to identify which experiments to run, we engineer systems where an AI agent proposes different experiment variants, marketers select them to test against the control group, and the algorithm automatically implements the winning variant.
For customers, with the increased adoptions of LLMs in the customer discovery funnel, customers are increasingly looking for experiences that are quicker, more personalised & convenient than ever. We are transitioning away from the legacy digital funnel which relied heavily on manual search, filtering and a fixed customer journey toward an AI-driven orchestration layer which is able to deliver a ‘one-shot’ recommendation tailored to the customer’s unique context.
This means that the customer attention span is shorter and brand equity may become less relevant compared to speed of service. Aside from Generative Engine Optimisation (GEO), businesses should start thinking about their API strategy and how their relationship with customers changes when purchase decisions are outsourced and facilitated through agents and LLMs.
Much of this technology is nascent today because adoption is still limited, but I believe companies which innovate early will build a competitive edge in this new frontier.

