How Mondelēz Built Always-On Consumer Intelligence Across 29 Markets with AI and Human Research

“A senior stakeholder asked me a question in the morning. By end of day I had an answer grounded in real consumer data we’d already collected. No new study, no waiting.”

— Anna Citro, Innovation & Strategy Insight Manager, Europe, Mondelēz International

From the Discuss session at Quirks London, How Mondelēz Is Unlocking Always-On Consumer Intelligence with AI-Powered Research, London, 2026.

Watch the full session here.

Mondelēz International runs consumer research at a scale that most insights teams only plan for: 337 projects, 1,254 sessions, 29+ markets, over a decade. The scale matters less than what it enables. Because Discuss compounds across every project, Anna Citro, Innovation & Strategy Insight Manager for Chocolate in Europe, can answer a senior stakeholder’s question the same morning it arrives, not by launching new research, but by querying a decade of intelligence her team already built. This case study explains how Mondelēz got there.

TL;DR

• Mondelēz has run 337 projects and 1,254 sessions with Discuss across 29+ markets over 10 years, totaling $2.6M in engagement value
• Virtual Personas, built from prior consumer interviews, let the insights team answer same-day stakeholder questions without commissioning new studies
• AI-assisted analysis accelerated reporting while keeping Anna in control of synthesis and interpretation
• Multi-method flexibility on a single platform (live interviews, async shopalongs, ideation workshops) reduced costs compared to a traditional full-service agency model
• Every new project adds to a compounding intelligence base that grows more useful with time

About Mondelēz International

Mondelēz International is one of the world’s largest snacking companies, home to brands including Cadbury, Oreo, Milka, Toblerone, and Ritz. Its products reach billions of consumers across more than 150 countries. Understanding how those consumers snack, gift, and treat themselves is not a research function; it’s a strategic one.

Anna Citro, Innovation & Strategy Insight Manager for Chocolate in Europe, is responsible for sharing, gifting, and seasonal insights. Her work sits at the intersection of fast-moving innovation programs, senior stakeholder demands, and the challenge of making consumer understanding accessible across functions that don’t always have direct consumer access.

What problems were Mondelēz’s insights teams actually trying to solve?

Three pressures, and Anna was quick to note “not just me” when she heard the same themes repeated across sessions at Quirks London.

The first was speed. “We need to be faster than yesterday,” Anna said. “We need to be delivering insights and results to the team as soon as possible.” On cross-functional innovation projects with senior stakeholders involved, research had to keep pace with decisions, not catch up to them.

The second was budget. Spending less didn’t mean doing less. “Budgets are getting cut, but that doesn’t mean the team expects less of us,” Anna said. The goal was reaching more consumers across more markets without proportional cost increases — a pressure industry research like the GreenBook GRIT Report has documented across insights teams.

The third was activation. Mondelēz had accumulated significant consumer research over years. The harder problem was putting it to work. “We have so much knowledge around, and sometimes there is a struggle even in using the knowledge that we already have as a company,” Anna said. New research also needed to be built to last: shareable and accessible well beyond the insights team.

How did Mondelēz approach research as a program rather than a series of projects?

Anna’s answer to all three pressures was treating research as a program. A program adapts. It supports shifting briefs, new markets, and a range of methods without requiring a new vendor, a new setup, or a new brief every time.

The gifting innovation program Anna ran last year drew on shopalongs, product placement sessions, one-on-one interviews, and an ideation workshop, each designed to answer a different question at a different stage. Some sessions were fully moderated with a Discuss team handling reporting. Others Anna ran herself. The platform supported both without friction. “I like not being forced to go to one vendor for one thing specifically,” she said. “In certain projects, you just don’t know what you’re going to need.”

How does Discuss combine AI analysis and human research in one program?

This is where Discuss differs from AI-only research tools. Discuss is built on the principle that AI analysis and human expertise work best together, not as competing approaches, but as two parts of one integrated methodology.

AI-assisted analysis in Discuss surfaces themes, flags relevant quotes, and compresses the time between session completion and stakeholder-ready outputs. When Anna was leading reporting herself, the AI summaries served as a check on her own synthesis, surfacing patterns she could validate or challenge. When she relied on the full-service team, the same tools let her stay close to the research even when she couldn’t attend every session. “It really helped me in creating those snackable insights faster,” she said.

Live moderation served a different function. It gave cross-functional stakeholders direct access to consumers they’d never had before. “It was really nice when marketing, R&D, packaging, and even project management were jumping on the connects live,” Anna said. “It meant they were really feeling engaged with the project.” Teams that had never spoken to consumers directly were now returning to video clips weeks after the original fieldwork, during their own ideation sessions.

Discuss combines AI and human research because both produce what the other can’t. AI scales analysis and speeds up what a skilled researcher does. Human moderation, judgment, and expertise make those outputs worth acting on. The human-led research feeds back into the AI layer, making analysis and Virtual Personas more accurate over time.

What is Virtual Personas, and how did Mondelēz use it to answer same-day stakeholder questions?

Virtual Personas is a Discuss feature that builds AI-powered personas from a brand’s own proprietary research data. Every response the persona generates is traceable to source interviews. Insights teams use it to pressure-test hypotheses, explore positioning, and answer strategic questions before investing in new fieldwork.

When a senior stakeholder asked Anna whether “premium occasions” genuinely resonated with consumers, or whether it was internal language the business had invented, she had a hypothesis, but not enough to share upward with confidence. She used Virtual Personas, built from prior research with premium chocolate buyers, to pressure-test the question that same morning.

The response surfaced a nuance her team had sensed but hadn’t clearly articulated: premium occasions aren’t only tied to external events like birthdays and anniversaries. They’re also self-created: a bad day that calls for a treat, a personal milestone worth marking. By end of day, Anna had shared a response grounded in real consumer data with the stakeholder who asked.

“You’re not just going back to research,” she said. “You’re sharing something new, but based on real knowledge that we developed.”

That’s the difference between research as a filing system and research as a living intelligence resource.

How quickly does Discuss deliver consumer research outputs?

Speed showed up at every level of the program. The shopalong component of the gifting project was briefed on a Thursday. Edited consumer video clips were in stakeholders’ hands by Monday. “That kind of speed changes what’s possible in an innovation project,” Anna said.

For the ideation workshop, consumers recorded themselves shopping for gifts in-store using self-paced video activities, capturing real-time decision-making and the moment a product earns its place in a basket. Discuss delivered edited clips within a day. The clips stayed useful long after the session ended. “They would go back to those videos,” Anna noted, “and it would just help them for their own ideation process.”

AI-assisted reporting reduced the gap between session completion and shareable outputs. Virtual Personas answered strategic questions the same day they arrived, without new fieldwork.

What results has the Mondelēz and Discuss relationship produced over 10 years?

Across 337 projects, 1,254 sessions, and $2.6M in total engagement value, Discuss has supported Mondelēz’s global insights teams across product development, brand work, and innovation, spanning 29+ markets.

Cost compared to a single full-service agency model was meaningfully lower. “If I had done that with a single partner, like a regular core partner, the end cost would have been much higher,” Anna said. The savings came from using the right method for each research question and from handling more of the process in-house, not from cutting research quality.

Cross-functional engagement improved in ways the insights team hadn’t seen before. For the first time, marketing, R&D, packaging, regional, and local teams were joining consumer sessions live. They were returning to video clips during their own work. Closer to consumers than they’d been, teams used that proximity to change how decisions got made. “The platform and the ease of usage made it really nice for the team to engage,” Anna said. “We got a great turnaround of those partners joining us.”

Every project Mondelēz runs with Discuss adds to an intelligence base that can be queried, revisited, and built on. Past research informs new hypotheses. Virtual Personas are built from real interviews. The intelligence grows more useful with every study, which is what an always-on research program is designed to do.

Why does Discuss work for enterprise consumer research programs at global scale?

Discuss is an always-on market insights solution that combines live human interviews, AI-moderated research, asynchronous video studies, Virtual Personas, and quantitative surveys in one solution. It operates across 150+ countries and supports 50+ languages.

Discuss is a Forrester Wave Leader for Experience Research Platforms, Q1 2026. Forrester gave Discuss the highest possible score in live interviews, AI-powered research methods, participant quality, participant recruitment, data analysis, multiple languages, and platform usability, among other criteria.

For a global program like Mondelēz’s, the all-in-one platform matters for a concrete reason: every new project adds to the same base of consumer intelligence rather than starting over in a new tool. Switching between full-service and self-service doesn’t require a new vendor. AI analysis and human research stay connected in the same environment, so each informs the other.

Mondelēz isn’t alone in this approach. See how Danone built a global consumer closeness program, how Reckitt speaks to global consumers at scale, and how agencies like buzzback and Walnut Unlimited deliver customer voices quickly, or browse all Discuss customer stories.

FAQ

Ready to see what this looks like for your organization? Talk to us or download the Forrester Wave™: Experience Research Platforms, Q1 2026 — the independent evaluation that named Discuss a Leader in the category.