Mid-Year Predictions Check-In: When the Future Arrives Faster Than Expected

In December 2024, I published "My 8 predictions for our industry in 2025," including a controversial bet that Omnicom's acquisition of IPG wouldn't complete. Six months later, as we hit the year's midpoint, it's time for the reckoning every predictor dreads: were we right, wrong, or did reality prove more complex than anyone anticipated?

There's something humbling about reviewing your own predictions. Unlike the pundits who quietly hope their failed forecasts disappear into the internet's memory hole, I believe there's value in the forensic examination of what we got right, what we got wrong, and perhaps most importantly, what we completely failed to anticipate.

The exercise has taught me something uncomfortable: in our industry, being directionally correct but temporally wrong can be worse than being completely wrong. When you underestimate the speed of change, you risk being prepared for tomorrow's challenges with yesterday's timeline.

But first, let me acknowledge what this review isn't: it's not a victory lap. Half my predictions proved too conservative, and one seems almost laughably premature. Instead, this is an attempt to extract strategic value from the gap between expectation and reality – because that gap reveals where the real opportunities and threats lie.

1. Bye AI PoC, hello Scale — On Track

This one feels almost quaint now. Six months ago, I predicted 2024's "Year of the POC" would give way to scaled AI implementation. What I underestimated was the velocity. We've moved beyond tentative scaling to wholesale workflow transformation. The compression of content supply chains isn't happening gradually – it's happening in quarterly leaps.

Yet I missed something crucial: the human cost. As we've scaled AI adoption, we've simultaneously hit what I now call "the capability ceiling" – the point where human adaptation can't keep pace with technological advancement. The bottleneck isn't the technology anymore; it's our ability to evolve our skills, processes, and mindsets quickly enough.

3. Set Creativity Free — On Track (and accelerating)

The bifurcation I predicted between one-stop integrated shops and specialised boutiques has crystallised faster than expected. Publicis's market share gains demonstrate the power of true integration, while the "muddy middle" – WPP, Havas, Dentsu – finds itself increasingly squeezed. The holding companies that can't decide whether they're platforms or collections of agencies are paying the price.

What's fascinating is how this split mirrors broader market dynamics. Clients increasingly want either complete solutions or best-in-class specialists – the middle ground of "pretty good at everything" has become commercially untenable. The agencies thriving are those that made clear strategic choices rather than trying to hedge their bets.

8. While Rome Burns — On Track

The Omnicom-IPG merger is proceeding as predicted, passing regulatory scrutiny and investor approval. What I called correctly was the market's acceptance of the "rust under the panels" – investors seem willing to overlook integration risks for the promise of scale. This sets up the industry's new structure: the Big 3 (Publicis, Accenture, Deloitte) versus everyone else - will Omnicom credibly join the big three or stay in the middle, distracted by integration and reduction?

My prediction about "$750 million in annualised synergies" being "gains for shareholders, not clients" has proven prescient. The integration costs are mounting, with leadership distraction visible in client service delivery. Several major accounts have been put in review not because of performance issues, but because clients are uncertain about service continuity during the merger process. IPG accelerated taking more than 1/3 of these costs themselves pre-merger due to the urgency to act.

The human toll is becoming apparent. IPG's latest restructuring announcement affects another 2,000 positions, while Omnicom has quietly implemented hiring freezes across non-essential functions. The "synergies" are people, and those people are reading the writing on the wall.

2. Play the Orchestra — More Urgent Than Expected

I predicted orchestration would become the most valuable agency skill. Six months in, it's become the most desperately needed. The proliferation of AI point solutions isn't creating the seamless "better, faster, cheaper" I envisioned – it's creating new forms of complexity. Most agencies are adding tools without removing friction, making marketing more complicated rather than more flowing.

Here's what I didn't anticipate: the orchestration challenge isn't just technical – it's cultural. Agencies are excellent at acquiring new capabilities but terrible at retiring old ones. They're layering AI content generation on top of existing approval processes, adding predictive analytics while maintaining traditional planning cycles, implementing automated optimisation while preserving manual oversight structures.

The result? What I call "complexity debt" – organisations where the promise of AI efficiency is undermined by the persistence of pre-AI bureaucracy. The agencies winning are those brave enough to eliminate entire workflow steps, not just accelerate them.

5. Glass Boxes, not Black Ones — More Critical Than Expected

Meta's Q1 earnings call revealed Zuckerberg's vision for capturing "a meaningfully larger share of global GDP" through complete platform control. His description of future advertising – "you don't need any creative, you don't need any targeting demographic, you don't need any measurement, except to be able to read the results that we spit out" – makes my prediction about platform transparency seem almost naive. It's not just a block box, it's also becoming titanium.

What I underestimated was the brazenness of the platform power grab. Where I predicted platforms would "innovate to disintermediate media agencies," the reality is more aggressive: they're attempting to eliminate the entire marketing value chain except budget allocation and results reporting.

The implications extend beyond media. When Zuckerberg describes a world where advertisers "connect to your bank account" and let Meta handle everything else, he's describing the death of marketing as a profession. Not evolution – extinction.

Yet here's what gives me hope: client pushback is building. Several major advertisers have privately expressed concern about this trajectory. The challenge is coordinating resistance when individual defection seems economically irrational.

This is where industry leadership becomes critical. The holding companies, for all their flaws, may be the only entities with sufficient scale to negotiate platform accountability. If they can't deliver transparent, auditable performance measurement, we'll have traded marketing sophistication for algorithmic convenience – and future CMOs will wonder how we gave away so much control.

4. Search becomes Find — Accelerated Beyond Prediction

I predicted LLMs would become our first place to find information. What I didn't foresee was how quickly this would devastate traditional search traffic. Google's AI Overview rollout has reportedly cut Reddit's natural search traffic in half since December. Zero-click searches aren't gradually increasing – they're exploding. The transformation from search to find is happening in months, not years.

The speed has caught even Google off-guard. Internal documents suggest they expected a more gradual transition, allowing publishers time to adapt their monetisation strategies. Instead, the change has been so rapid that major content platforms are experiencing what one executive called "traffic cliff events" – sudden, dramatic drops in organic reach that threaten their entire business models.

This creates a fascinating paradox for brands. Traditional SEO strategy focused on ranking for search terms that drove traffic to owned properties. But if AI systems answer questions directly, the value of "traffic" diminishes while the value of being cited as a source increases exponentially.

I'm seeing smart brands pivot from "content that ranks" to "content that teaches AI systems." This means optimising for how LLMs perceive and cite your information rather than how search algorithms rank your pages. It's a completely different discipline that most marketing teams aren't prepared for.

What worries me is how few marketing leaders understand the implications. If consumers stop visiting websites, traditional conversion funnels collapse. If product discovery happens through AI conversation, traditional advertising becomes irrelevant. We're not just watching search evolve – we're watching the entire customer journey transform.

6. Repeat after me: Multimodal AI Agents — Rebranded but Right

I predicted "Multimodal" would be Cannes 2025's buzzword. Instead, "Agentic" dominated the Croisette. But scratch beneath the terminology and it's the same fundamental shift: marketing to humans is giving way to agents marketing to agents. Despite the industry buzz, there's still widespread misunderstanding about how this agent-to-agent commerce will fundamentally reshape the marketing value chain.

The linguistic evolution is telling. "Multimodal" felt technical and intimidating – it required explanation. "Agentic" feels actionable and immediate, even if most people using the term can't define it precisely. This semantic shift reveals something about our industry's relationship with complex technology: we prefer concepts that sound accessible even when the underlying reality is anything but.

But here's what concerns me about the agentic AI conversation at Cannes: most discussions focused on how agents could execute existing marketing strategies more efficiently. Very few addressed how agent-to-agent commerce will require completely different strategies.

Consider this scenario: when AI shopping agents become commonplace, they won't browse websites or respond to traditional advertising. They'll evaluate products based on structured data, peer agent recommendations, and algorithmic trust scores. The entire concept of "brand awareness" becomes irrelevant when the decision-maker isn't human.

I've started asking clients a simple question: "What happens to your marketing strategy when the customer never sees your marketing?" The silence is usually deafening. We're preparing for a world where marketing happens between machines, but we're still thinking about human psychology and emotional connection.

The agencies that survive this transition will be those that understand they're not in the persuasion business anymore – they're in the information architecture business. They need to make their brands discoverable, comparable, and trustworthy to algorithms rather than humans.

7. Blockchain's Back, Baby — Not Yet, But Soon

This remains my most premature prediction. Blockchain technology hasn't made its anticipated comeback – yet. But as we move toward agent-to-agent marketing transactions, the need for transparent attribution and verification will become undeniable. When marketing becomes invisible investment in AI agents engaging with one another, blockchain won't be a nice-to-have for attribution – it'll be essential infrastructure.

I'll admit this one makes me look like a crypto enthusiast, which I'm not. My prediction wasn't based on blockchain's speculative potential but on its practical necessity for a world where trust must be automated.

Here's the problem we're rapidly approaching: when AI agents make purchasing decisions on behalf of humans, how do we verify that those decisions weren't manipulated by other AI agents? When marketing becomes algorithmic influence rather than human persuasion, we need algorithmic verification of that influence.

Traditional attribution relies on human-auditable processes. But when agents operate at millisecond speeds across millions of transactions, human auditing becomes impossible. We need systems that can verify agent behaviour automatically, transparently, and immutably – which is precisely what blockchain technology provides.

I'm starting to see early signals that this prediction might be just delayed rather than wrong. Several major tech companies have quietly filed patents for blockchain-based advertising verification systems. Google has been testing cryptographic verification for some search results. Amazon's agent-to-agent commerce pilot includes blockchain audit trails.

The catalyst won't be marketing executives suddenly embracing blockchain. It will be regulatory pressure for transparent, verifiable AI agent behaviour. When governments demand auditable trails for algorithmic decision-making, blockchain will shift from crypto speculation to compliance necessity.

My biggest miss wasn't about direction but velocity. I consistently underestimated how quickly these changes would accelerate. The AI POC-to-scale transition happened in quarters, not years. Search disruption is measuring in months, not years. Platform consolidation is happening now, not gradually.

This speed creates new challenges I didn't fully anticipate. The human bottleneck in AI adoption has become more severe because technology is advancing faster than organisational capability to absorb it. The orchestration challenge is more urgent because point solutions are proliferating faster than integration strategies.

But there's a deeper issue I missed entirely: the psychological impact of exponential change on marketing professionals. When I wrote about AI adoption, I focused on technical and strategic challenges. I didn't consider the human cost of constant adaptation.

I'm now seeing what I call "change fatigue" across the industry. Marketing teams that successfully implemented AI workflows in Q1 are being asked to integrate new platforms in Q2, adopt agentic tools in Q3, and prepare for agent-commerce in Q4. The pace isn't just technically challenging – it's emotionally exhausting.

This creates a curious dynamic: the organisations most successful at early AI adoption are sometimes the most resistant to continued evolution because their teams are burned out from constant change. Meanwhile, companies that were slow to start are now leapfrogging ahead because they can implement more advanced, integrated solutions rather than layering new tools onto legacy processes.

The velocity also revealed something uncomfortable about industry expertise. Traditional marketing knowledge – developed over decades of stable media consumption patterns – became obsolete in months rather than years. The experts who built careers on understanding consumer behaviour suddenly found themselves trying to understand AI agent behaviour instead.

What's particularly striking is how this speed differential has created new forms of competitive advantage. The companies thriving aren't necessarily those with the best strategies, but those with the highest adaptation velocity. In a world where being six months behind means being irrelevant, agility trumps excellence.

One theme that emerged across all my predictions is that the challenge isn't technological – it's human and organisational. AI tools work remarkably well. Platform capabilities are sophisticated. The systems exist to create the "better, faster, cheaper" marketing I predicted.

What doesn't exist is the institutional capability to implement these tools effectively. We have cutting-edge technology being deployed by organisations designed for pre-digital marketing. It's like installing a Formula 1 engine in a horse-drawn carriage.

The companies succeeding in this environment share three characteristics I didn't anticipate:

Ruthless Simplification: They're not just adding new tools – they're aggressively removing old processes. Every new capability must eliminate a legacy workflow, not augment it.

Behavioural Metrics: They measure learning velocity, adaptation speed, and change resilience as seriously as they measure traditional KPIs. They've recognised that in exponential change environments, capability development is more important than capability deployment.

Psychological Safety: They've created environments where admitting ignorance about new technologies is rewarded rather than punished. The teams adapting fastest are those where senior leaders model vulnerability about what they don't know.

This last point particularly surprised me. In my predictions, I focused on technical and strategic considerations. I didn't appreciate how much of transformation success depends on creating cultures where people feel safe to experiment, fail, and learn quickly.

Based on what we've learned in the first half, here's what I'm watching for in H2:

Which agencies will crack the code on making multiple AI tools work together seamlessly rather than just accumulating more of them? I'm particularly watching for agencies that start retiring legacy processes rather than just accelerating them. The winners will be those that eliminate entire workflow layers, not those that optimise existing ones.

Watch for: Agencies announcing significant staff reductions accompanied by capability increases. The counterintuitive signal of success will be smaller teams delivering more output.

How quickly will brands begin optimizing for AI agent discovery rather than human search behavior? This shift will start with B2B companies where purchase decisions are more algorithmic and less emotional.

Watch for: Job postings for "Agent Experience Specialists" or similar roles. The early signals will be new professional categories focused on algorithmic rather than human interaction.

Which marketing skills become valuable and which become obsolete as transformation accelerates? I'm particularly interested in how traditional creative roles evolve in an AI-augmented environment.

Watch for: Changes in agency hiring patterns and compensation structures. The leading indicators will be which roles see salary increases and which see stagnant or declining compensation.

How will accelerating change affect talent retention and acquisition in marketing? I suspect we'll see significant movement as professionals seek teams better adapted to the new environment.

Watch for: Senior marketing talent moving from traditional agencies to tech companies or starting independent ventures. The brain drain from legacy organisations to adaptive ones will accelerate (and yes, this includes watching out for what I do next.... stealth mode continues....)

Amid all this acceleration, there's something oddly comforting about what remains constant: human nature. Despite AI agents and algorithmic marketing, people still want to feel understood, valued, and connected to brands that reflect their values.

The successful companies aren't those that have eliminated human elements, but those that have used technology to enable more authentic human connections. AI content generation works best when it amplifies human creativity rather than replacing it. Agent-driven commerce succeeds when it serves human preferences more effectively than human-driven processes.

This suggests that the ultimate winners in our industry's transformation won't be the most technically sophisticated organisations, but those that maintain human insight while leveraging technological capability. The goal isn't to become more like machines, but to become more effectively human.

That's perhaps the most important lesson from this prediction review: in a world of exponential technological change, sustainable competitive advantage comes from the things that don't change – human creativity, authentic connection, and the ability to solve meaningful problems for real people.

The technology will continue accelerating. The platforms will continue consolidating. The pace will continue intensifying. But the fundamental purpose of marketing – connecting human needs with valuable solutions – remains wonderfully, reassuringly constant.

The challenge with predictions in our industry isn't being wrong – it's being right too slowly. The future has a habit of arriving faster than we anticipate, making even accurate predictions feel conservative.

What remains constant isn't the pace of change but its direction: toward greater integration, more automation, increased platform power, and the need for transparent, orchestrated solutions that create flow rather than friction.

The question for the second half of 2025 isn't whether these trends will continue – it's whether we can adapt quickly enough to benefit from them rather than be disrupted by them.

How are your own 2025 predictions holding up? What did you get right, wrong, or fail to anticipate the speed of?

Brandflow is written by Justin Billingsley, who has spent his career on all three sides of the industry's table: senior client, global agency leader, technology founder. First published 1 July 2025 in the Brandflow newsletter on LinkedIn.