This Month in Martech

AnalyticsDev.net 2027 Conference

17 March 2027

Copenhagen

AI Is Becoming Part of the Marketing Operating Layer

October 2026

From conversational analytics and AI coworkers to governed customer context and agent-ready content, September’s developments suggest MarTech is moving beyond adding AI features and towards rebuilding the marketing technology stack around agents.

For much of the past two years, the MarTech industry’s AI story has centred on adding copilots, assistants and generative capabilities to existing platforms.

September suggests something more fundamental is happening.

Across customer data, analytics, content management, experimentation and enterprise data, technology companies are increasingly designing platforms for a world in which AI agents do not simply answer questions. They retrieve data, interpret business context, configure systems and increasingly perform work.

That raises a more consequential question for marketing leaders:

What infrastructure does an AI-powered marketing organisation actually need?

Customer data is becoming context for AI

Tealium provides a good illustration of the transition.

Its recent releases include Tealium Studio, a natural-language interface for interrogating server-side configurations, alongside Configuration MCP, which enables AI agents to read and update configurations. Tealium has also renamed Moments API as Context API, reinforcing the idea of providing customer context to downstream systems.

September brought further infrastructure developments, including general availability of the iQ Publish API, an Amplitude Cohorts data source and continued expansion of server-side connectors.

Source: Tealium release notes

The strategic question is larger than any individual release.

Connecting an LLM to data is becoming relatively straightforward. Giving an AI system reliable customer identity, consent, behavioural history and organisational definitions is considerably harder.

This could become one of the defining architectural questions in MarTech: which technology owns the context that allows an AI agent to understand the customer?

CDPs, data platforms, analytics systems and marketing applications increasingly have overlapping answers.

Marketing analytics is becoming conversational

A similar transition is happening in marketing analytics.

Supermetrics introduced Business Context for its MCP server in September. Teams can save reporting rules, account context and preferences that subsequently become available to AI conversations. Its SuperAI capability can also answer natural-language questions using live marketing data.

Source: Supermetrics September 2026 product updates

Funnel is moving in a similar direction. Its MCP server can now connect to Microsoft Copilot Studio, enabling agents to query Funnel workspace data and field definitions through read-only tools.

Source: Funnel September 2026 release notes

The interesting development isn’t simply natural-language querying.

Traditional marketing analytics has generally followed a familiar sequence:

collect → transform → dashboard → interpret → decide

AI potentially compresses that workflow into something closer to:

ask → investigate → understand → act

A marketer might start with a question such as: Why did paid acquisition performance deteriorate last week?

The system can determine which data it needs, retrieve it, analyse it and explain the result.

Dashboards are unlikely to disappear, but their position as the primary interface between marketers and their data increasingly looks open to challenge.

Analytics is moving into the tools where decisions happen

Contentsquare offers another variation on the same idea.

Its September releases allow users to query Contentsquare data directly from Figma and Dust through its MCP server. Teams can ask questions about metrics, journeys and funnels without leaving the environment where they are working.

The company has also expanded Sense Analyst, including the ability to add business context and deliver insights through email, Teams and Slack.

Source: Contentsquare 2026 releases and updates

This points towards an important behavioural change.

For years, analytics vendors have tried to persuade more employees to visit analytics platforms.

AI may invert the model.

Rather than bringing people to analytics, analytics increasingly comes to the places where people are already making decisions.

Designers can receive behavioural evidence while designing. Marketers can investigate performance inside an AI workspace. Executives can interrogate customer intelligence through familiar productivity environments.

The analytics platform remains important, but its interface becomes distributed.

Enterprise data platforms want to govern the agent layer

Snowflake is approaching the same transformation from the enterprise data side.

Its September AI Pulse highlighted AI Gateway, providing model routing, cost governance and centralised MCP management, alongside Semantic Studio, designed to provide trusted semantic context at scale. Snowflake also demonstrated further automation capabilities through Coco and governance developments around Cowork.

Source: Snowflake AI Pulse, September 2026

The strategic significance can easily be lost among the individual features.

Enterprise AI needs more than access to data. It needs to understand what that data means.

A company’s definition of an active customer, qualified lead, profitable campaign or valuable audience isn’t contained in raw numbers alone.

Semantic models and business context therefore become important infrastructure.

The next competitive battle may consequently be less about who stores the most data and more about who provides AI with the most trustworthy understanding of that data.

AI coworkers are moving closer to actual work

Optimizely represents another stage in the transition.

Its Virtual Teammates proposition explicitly moves beyond the conventional chatbot model. Optimizely describes them as systems designed to act, execute and deliver rather than simply waiting for individual prompts.

Source: Optimizely: Introducing Virtual Teammates

This is quite different from asking an AI assistant to produce five alternative headlines.

The technology is beginning to participate in the operating workflow itself.

That creates a new management question:

Which marketing activities should AI be permitted to perform rather than merely recommend?

Permissions, approvals, brand governance and accountability consequently become as important as model capability.

That question extends beyond MarTech vendors themselves. Microsoft’s September Copilot Studio roadmap, for example, includes the ability for organisations to require explicit human approval before agents execute specified tools.

Source: Microsoft Copilot Studio release plan

The movement towards autonomous marketing therefore appears likely to be accompanied by a parallel movement towards more explicit governance.

Brands increasingly have two audiences

Perhaps one of the month’s more provocative propositions comes from Contentstack.

Its ContentCon 2026 positioning argues that brands increasingly need to design digital experiences for two audiences: humans and AI agents. The company is encouraging organisations to rethink content, governance and personalisation for an environment in which AI systems increasingly mediate digital experiences.

Source: ContentCon 2026: Every Brand Now Has Two Audiences, Humans and Agents

This extends the current discussion around AI search.

A brand’s content increasingly needs to do more than persuade a human visitor or rank in a conventional search engine. It may also need to provide sufficient structured context for an AI system to understand products, services, policies and brand positioning accurately.

The question for brand leaders becomes:

Does an AI system understand your brand well enough to represent it when a customer asks for advice?

That could elevate areas such as content architecture, product information and structured brand data from technical considerations into strategic brand issues.

AI-powered customer experience moves into enterprise deployment

Adobe and Jet2 provided a useful real-world example this month.

The companies announced a multi-year strategic partnership built around Adobe CX Enterprise and agentic AI. Jet2 plans to use the technology to deliver more personalised recommendations, offers and support across the holiday journey.

The partnership also establishes a Jet2 x Adobe Customer Experience Lab, bringing Adobe engineers and Jet2 teams together to develop new AI-powered customer services.

Source: Adobe and Jet2 agentic AI partnership announcement

This matters because it moves agentic CX beyond vendor demonstrations into organisational implementation.

Instead of asking only what AI functionality a platform contains, marketing leaders increasingly need to consider how people, technology, customer data and operating processes change together.

The bigger picture: MarTech is developing a new operating layer

Taken together, September’s developments suggest that the MarTech conversation is shifting.

The first phase of generative AI was largely about creation:

Write the email. Produce the image. Summarise the report.

The emerging phase is about context and action.

AI needs to know what a customer means.

It needs reliable marketing and behavioural data.

It needs definitions of business metrics.

It needs permissions.

It needs privacy and governance rules.

It needs access to the systems where work actually takes place.

And once it has those things, organisations must decide what it is permitted to do.

This makes several previously technical subjects considerably more strategic: data governance, semantic layers, tracking plans, APIs, MCP, consent architecture and customer identity.

They are becoming part of the infrastructure through which AI understands and interacts with the business.

What marketing leaders should watch next

Three questions are becoming particularly important.

Who owns the context layer?

CDPs, enterprise data platforms, analytics providers and marketing applications all want to supply trusted context to AI. Traditional MarTech category boundaries could become increasingly blurred.

What happens to the dashboard?

As conversational analytics improves, marketers may spend less time navigating predefined reports and more time interrogating data through AI. The dashboard becomes one interface among many rather than necessarily the destination.

Where should human approval remain mandatory?

AI systems are progressing from recommending actions towards performing them. Organisations will need clearer policies governing what agents can analyse, create, configure, publish and change.

The most important MarTech question may therefore no longer be simply:

“Where are we using AI?”

It may increasingly become:

“What does our marketing technology stack need to look like when AI becomes one of its users?”

Sources

Editorial note

This Month in MarTech is an editorial briefing from the Brand Leadership Community examining significant developments across marketing technology, analytics, customer data, AI, digital experience and related fields. Companies and developments are selected for their relevance to the wider MarTech landscape and should not be interpreted as endorsements of particular vendors or products.

About AI and this article

This Month in MarTech is produced with AI assistance using a consistent Brand Leadership Community research and editorial framework. AI supports research, source comparison, synthesis and identification of wider industry patterns. The briefing draws on current public sources and BLC’s knowledge of the MarTech ecosystem. Editorial conclusions are intended to help readers understand emerging developments rather than recommend particular technology providers.