The Biggest Takeaways from Martech for 2026 (And Why Most Teams Are Still Experimenting)

I've just finished reading Scott Brinker and Frans Riemersma's Martech for 2026 report—all 122 pages of it—and I need to talk about what's actually happening with AI in marketing. Not the hype, not the predictions, but the real operational reality that hundreds of marketing leaders are navigating right now.

The report surveyed marketing operations professionals from around the world, and the data reveals something fascinating: we're in a massive transition period where AI agents have moved from science fiction to everyday tools, yet most teams are still figuring out how to operationalise them properly.

AI Agents Are Real (But Not How You Think)

Let's start with the headline number: 90.3% of marketing organisations now use AI agents in some capacity. That's not a typo. Nine out of ten teams have AI agents doing something in their marketing operation.

But here's the reality check: only 23.3% have agents in full production. The vast majority are still experimenting, piloting, or running agents in carefully constrained workflows. This gap between adoption and operationalisation represents the biggest opportunity for competitive advantage in 2026.

Where AI Agents Actually Work

Most AI agents today work behind the scenes. They're producing content (68.9% of teams), discovering audiences (40.8%), conducting competitive analysis (35.9%), and enriching data. They're "assistants" rather than autonomous systems—80.6% operate with a human in the loop.

"Things become agentic when you combine the ability of AI models to reason with tools and perform actions in the external world. So things become agentic when you combine tools like reverse ETL with the LLMs that can understand English intent." — Tejas Manohar, Co-CEO of Hightouch

The customer-facing applications are growing too: 54.4% use AI-powered chatbots, and 21.4% have deployed AI SDRs. But the real action is in the backstage operations where AI agents are creating meaningful efficiency gains whilst freeing marketers to focus on strategy.

The Real Problem Isn't AI—It's Data

If there's one finding that should keep marketing directors up at night, it's this: 56.3% of respondents said poor data quality was their biggest AI implementation challenge. Missing data, stale data, inconsistent data across systems.

I've seen this first-hand with clients. Last quarter, we worked with a B2B SaaS company that was excited to implement AI-powered lead scoring. They'd invested in a sophisticated agent that could analyse prospect behaviour and predict conversion likelihood. But when we dug into their CRM data, we found 40% of their records had incomplete information—missing job titles, outdated company data, duplicate entries. The AI agent was making decisions, sure, but it was essentially guessing based on incomplete information. We spent three weeks cleaning their data before we could even properly test the AI system. The lesson? You can't automate bad data into good decisions.

The survey found that 52.4% struggle with organisational and process readiness, and 50.5% face integration friction. Success with AI isn't about having the fanciest models—it's about having clean, accessible, well-structured data. Internal data sources being used include CRM and CDP profiles (61.2%), brand assets and content libraries (61.2%), and emails and messaging (48.5%).

Context Engineering: The New Core Competency

Here's a term you need to know: context engineering. It's replacing prompt engineering as the critical skill for AI implementation. Context engineering is about delivering the right data, content, functionality, and instructions to AI agents at the right moment.

The Context Engineering Framework:

Teams are integrating context through custom-built solutions (56.3%), pre-built platform capabilities (47.6%), and iPaaS platforms (40.8%). The Model Context Protocol (MCP) is gaining traction, with 48.5% using it in AI assistants and 27.2% in iPaaS or agents.

The Factory and The Laboratory

One of the most useful frameworks from the report is thinking about your martech stack as having two distinct roles: The Factory and The Laboratory. The Factory protects current revenue. It's your production systems—the tech stack that runs your campaigns, manages your database, processes transactions, and keeps the lights on.

The Laboratory creates future revenue. It's where you experiment with AI agents, test new approaches, and discover what's possible. This is probabilistic work where AI agents reason, adapt, and sometimes surprise you. Speed of learning matters more than perfection here.

The Hybrid Stack Reality

Most teams (85.4%) are using AI to enhance existing functionality, not replace it. Only 30.1% are using AI to replace existing capabilities entirely. The hybrid approach—blending deterministic workflows with probabilistic AI—is where the real sophistication lies.

"I think the biggest mistake that companies are making with gen AI when they're deploying is they're thinking about efficiency. They're not thinking about top-line growth. What gen AI unlocks is how do I do really personalised, highly targeted campaigns for every single one of my buyers? That's how you drive growth." — Rafael Flores, Chief Product Officer at Treasure Data

Despite all the talk about AI replacing traditional martech, the data shows something different. Your email platform doesn't need AI to reliably send messages to segmented lists. But AI can dramatically improve how you write subject lines, personalise content, or decide optimal send times.

Customer-Side Agents Are the Real Disruption

Here's what should fundamentally change your 2026 strategy: 50% of consumers already use AI-powered search today. McKinsey projects that $750 billion of consumer spend will flow through AI-powered search by 2028.

Buyers are using ChatGPT, Claude, Gemini, and Perplexity to research products and services. They're conducting entire evaluation processes inside these AI assistants before ever clicking through to your website. When they do arrive on your site, they're more ready to buy than before—but you have far less visibility into their journey.

Making Your Content AI-Discoverable

Only 63.1% of survey respondents are publishing AI-optimised content (structured Q&A, schema markup), and just 13.6% are measuring AI inclusion rate and agent-referred conversion. There's a massive opportunity here for teams who move quickly.

Five Actions to Support Customer-Side AI Agents:

1. Publish AI-optimised content with structured Q&A and schema markup

2. Create machine-readable feeds (JSON/CSV) for product and pricing data

3. Provide an MCP (Model Context Protocol) server for customer use

4. Publish an llms.txt or agents.json file

5. Expose deep links and APIs for customer-driven AI agents

The Mid-Market Opportunity

Intuit Mailchimp's research in the report reveals something important: 39% of mid-market marketers feel they don't have the knowledge or skills to embrace AI. But mid-market teams often have advantages that large enterprises lack—less bureaucracy, faster decision-making, and closer connections between marketing and product teams.

You don't need AI specialists. You need internal champions who create space for experimentation with clear use cases tied to specific business goals. The teams figuring this out aren't hiring data scientists—they're empowering their existing marketing ops people to experiment systematically.

What This Means for Your 2026 Planning

If you're planning your 2026 marketing technology strategy, here are three priorities to focus on. First, fix your data infrastructure. You can't effectively deploy AI agents without clean, accessible, well-governed data. If you're experiencing the data quality challenges that 56.3% of teams report, that's your starting point.

Second, create space for both The Factory and The Laboratory. Protect your production systems whilst carving out resources—time, budget, and people—for experimentation. The teams that figure out AI first will have learned through dozens of small experiments, not one big bet.

Third, prepare for customer-side AI disruption. Start publishing AI-optimised content now. Begin measuring AI inclusion rates. Understand that buyer journeys are changing faster than your attribution models can track. The traditional funnel tracking we've relied on for two decades is being fundamentally disrupted.

The Change Agent Is You

Here's perhaps the most important insight from the report: marketing operations professionals aren't being replaced by AI—they're entering a golden age. As AI agents proliferate, someone needs to vet them, deploy them, integrate them, and train others to use them effectively.

The challenge isn't technical. It's organisational. Success requires cross-functional alignment between marketing, IT, data teams, and legal. It requires balancing the efficiency gains that CFOs want with the growth experiments that drive future revenue. Marketing ops sits at the centre of this transformation.

Looking Ahead

We're in a transition period. AI agents are real, they're here, and they're being used by the vast majority of marketing teams. But we're still in the experimentation phase—figuring out what works, what doesn't, and how to operationalise the wins.

The teams that will win in 2026 aren't the ones with the fanciest AI tools. They're the ones who've sorted out their data, created space for systematic experimentation, and prepared for the fundamental shift in how buyers research and evaluate solutions.

The Martech for 2026 report is worth reading in full—it's packed with data, frameworks, and real-world examples that I couldn't cover here. But if you take away one thing, let it be this: the opportunity isn't in the AI itself. It's in how thoughtfully you integrate it into your existing operations whilst preparing for a market where AI agents are increasingly central to how buyers make decisions.