The Distillery #20: Meta's Brazilian Standoff & Google’s New Agentic Commerce Protocol

January 2026

Telecoms & Connectivity

Juniper Research

Analyst House

In this edition:

• WhatsApp faces a regulatory showdown in Brazil as authorities challenge Meta’s new chatbot restrictions.
• Google unveils its Universal Commerce Protocol, setting the stage for AI agents to take over the shopping journey.
• Launching an MVNO might be easier than ever, but here are five things to consider beforehand.
• Our latest insights on digital wallets, emerging tech, payments trends, and physical AI.

TELECOMS & CONNECTIVITY

Brazil Challenges Meta’s WhatsApp AI Restrictions

Brazil’s antitrust authority CADE has opened an investigation into Meta’s revised WhatsApp Business terms, which would block general-purpose AI chatbots from operating on the platform.

The policy is due to come into force on 15 January 2026, with Meta arguing that third-party bots generate unsustainable message volumes that strain its infrastructure. Brazil’s intervention follows similar scrutiny elsewhere: Italy secured an exemption earlier this week after its competition authority raised concerns under EU law, while the European Commission launched its own antitrust probe in December.

Distilled...

🟣 Brazil is too big to ignore. Our latest research shows 5.4 trillion OTT messages were sent in Brazil last year, and with an estimated 148 million WhatsApp users, the platform has become a core channel for business messaging, customer service, and automation. Policy changes in Brazil therefore have outsized implications for how enterprises deploy AI-driven engagement at scale.

🟣 Infrastructure constraints are becoming a competition issue. Meta’s argument that general-purpose AI chatbots overwhelm its systems may be technically valid at Brazil’s scale, but it also underscores how platform capacity decisions can shape which AI models are allowed to compete. As WhatsApp becomes critical infrastructure for business messaging, these technical limits are increasingly being assessed through a regulatory lens.

🟣 Market-by-market enforcement raises strategic risks. Italy’s exemption signals that regulators are willing to challenge blanket platform policies. If similar interventions emerge across other WhatsApp strongholds such as India, Mexico, and Indonesia, Meta risks operating under fragmented rules, while AI vendors and enterprises face uncertainty over where automation strategies can be deployed consistently.

GET THE WHITEPAPER

FINTECH & PAYMENTS

Google Launches Universal Commerce Protocol for AI Shopping

Internet giant Google recently unveiled the Universal Commerce Protocol (UCP), a new open standard to allow AI agents to be more seamlessly involved in the shopping journey; including discovery, purchase, and customer support. 

The announcement also includes AI-optimised tools for merchants, including branded shopping agents, as well as optimisations for ads to work better with AI discovery. The UCP was developed with key partners including Shopify, Etsy, Wayfair, and Target, with more than 20 other companies endorsing its launch.

Distilled…

🟣 By setting shared ground rules and integrations for agentic commerce, the UCP enables the next phase of exploration. With eCommerce sites set to handle 379.4 billion transactions this year, rising to 562.4 billion by 2030, the protocol gives retailers a way to move beyond theory and start testing agent-led experiences at real scale — and refine their strategies accordingly.

🟣 The framework has the backing to succeed, bringing together retailers and major payments players including Adyen, American Express, Mastercard, Stripe, and Visa. This early payments involvement reduces friction and makes large-scale rollout far more feasible.

🟣 Experimentation will be vital. Agentic commerce represents a fundamental shift in how transactions are initiated and completed; creating challenges existing systems were not designed for. For example, fraud management systems are based on human behaviour, so introducing AI agents means that fraud systems have to learn how to prevent fraud in new scenarios. The use of third-party APIs also creates additional cybersecurity risks, which will need to be mitigated.

GET THE WHITEPAPER

TELECOMS & CONNECTIVITY

So You Want to Launch Your Own MVNO?

MVNO in a Box platforms have made it easier than ever to launch a mobile service. Faster setup, lower upfront costs, and fewer technical hurdles mean enterprises no longer need deep telecoms expertise to get started.

But easier doesn’t mean automatic success. Removing the technical barrier to entry simply means the real battle has shifted to the brand and the bottom line. So to coincide with our new MVNO in a Box research, we’ve distilled five essential rules for the modern operator: from moving beyond simple data bundles as a differentiator, to leveraging AI for operational excellence rather than just marketing hype.

🟣 Not every business should launch an MVNO. First and foremost, putting your brand name on a mobile service isn’t a strategy. The MVNOs that work are tightly aligned with an existing customer base and a clear business goal; reducing churn, increasing engagement, or strengthening an ecosystem. Launching an MVNO for novelty alone is how brands end up damaging, rather than deepening, customer trust.

🟣 eSIM is ideal for testing, but hybrid models win. eSIM has dramatically lowered the barrier to entry, making it possible to launch, test demand, and refine propositions without heavy operational commitment. But physical SIMs still matter in many markets, especially among cost-conscious consumers and in regions with lower eSIM penetration. The strongest launches use eSIMs to prove the model, then scale with hybrid support once demand is clear.

🟣 Differentiation no longer comes from data allowances. In developed markets, cellular data growth is slowing as bigger bundles and cheaper plans lose their edge. Instead, successful MVNOs are embedding mobile services into wider loyalty and rewards schemes; turning connectivity into another touchpoint within an existing relationship, rather than a standalone product.

🟣 Superapps are the exception, not the rule. High-profile launches have made superapp-driven MVNOs look inevitable, but the reality is more restrained. Only a small number of enterprises have the scale, ambition, and operational appetite to pursue a true superapp strategy. For most, mobile works best as a complement to the core business, not the foundation of an entirely new one.

🟣 AI should make operations better, not promises bigger. AI is increasingly valuable across analytics, automation, and customer service, but overpromising autonomous or agentic capabilities risks harming the customer experience. The most effective uses of AI are also the least glamorous: better data analysis, smarter segmentation, and tools that help teams operate at scale without losing control.

GET THE WHITEPAPER

ICYMI

The Last Drop

We can’t fit all our work from the last fortnight in one issue, so here’s a hand-picked selection of blogs, infographics, and insights that deserve your attention.

📱 QR codes are set to dominate the digital wallet landscape, accounting for over 50% of all wallet transactions by 2028 as low-cost merchant adoption surges across emerging markets.

🔮 Your new tech roadmap is officially here. From neuromorphic computing to multi-agent systems, we’ve identified the top 10 emerging tech trends for 2026.

💰 If payments are your thing, we also broke down the three biggest trends for 2026; including the scaling of Pay by Bank in the UK, and how stablecoins are positioning themselves to rival traditional interbank settlement. 

🤖 Physical AI was everywhere at CES 2026, but what does the transition from "novelty robot" to "functional tool" really look like?

Subscribe

Contact Us