The hype around artificial intelligence may be heading into the “trough of disillusionment,” as Gartner calls it, but the technology continues to progress in big, disruptive ways — especially for martech. That is the nutshell version of the “Martech for 2025” report from Scott Brinker, editor of chiefmartec.com, and marketing technologist Frans Riemersma, released today.

“AI is reshaping marketing and martech,” they write. “And while we’re not prone to hyperbole, we do believe there will be significant real-world changes that marketers and marketing operations leaders will have to face with this technology in 2025.”

These changes go well beyond current use cases like content generation, personalization and knowledge management. It’s impossible to say what all the future use cases may be because AI’s ability to create “instant software” means they will be tailored to specific business needs. 

Here comes the martech hypertail

These new solutions, the “hypertail” according to the report, will be built not only by IT and marketers but also by AI agents. This means radical changes in the martech stack.

Dig deeper: 7 strategies for getting the most from your martech stack

“This may be the turning point where the number of commercial apps in the tech stack peaks and future growth of the stack — which overall we think could be exponential — will come from custom software, a cornucopia of custom apps, agents, and automations.”

Radical changes are already underway among martech vendors, changes that are upending the paradigm of how new companies compete with established ones.

Existing martech giants like Adobe, HubSpot, Microsoft, Salesforce and SAS have aggressively integrated new AI capabilities into their products, leveraging both generative and traditional machine learning. This has ignited a new battleground, reminiscent of the age-old startup vs. incumbent struggle.

However, a significant portion of AI-native startups aren’t directly challenging these incumbents. Instead, they’re innovating on the periphery, developing small, standalone tools that automate or enhance specific marketing tasks within existing platform ecosystems. These tools, powered by generative AI engines from OpenAI, Google, Anthropic, Meta and others, offer complementary solutions rather than direct competition.

Would you go all DIY?

One company, fintech giant Klarna, has gone so far as to ditch two major vendors in favor of DIY software. It is replacing Salesforce and Workday with its own custom CRM and HCM applications using AI and composable cloud services. 

While this may be an extreme move, as the report points out, “The fact that [it] is even conceivable is a testament to both the improved economics of custom development and the perceived business advantage of more tailored digital operations in the AI era.”

The other thing making this possible is the ability to tailor the AI using a business’s proprietary data and context-specific logic. The tailoring is primarily done one of three ways:

• Training one’s own model.
• Fine-tuning an existing model, and/or 
• Using retrieval augmented generation (RAG).

RAG, the most common method, looks up data from internal databases and feeds it into the prompts given to the LLM engine. The response is generated its response using that data as input, which augments the LLM’s generic knowledge. This has the bonus of providing further LLM guardrails because the RAG  can check or manipulate the LLM’s output.

Two other key points from the report:

• The importance of data strategy: Strong data strategies are fundamental for successful AI implementation. Cloud data warehouses are becoming essential for aggregating and orchestrating data. “The expression we’ve used for a long time… you don’t have an AI strategy if you don’t have a data strategy.”
• Composability is the key: Martech stacks must have modular, interconnected components for flexibility and adaptability. “With a composable approach, we can take the best parts and pieces and build out from there. You can compose a stack that is unique to your business needs and unique to your customer needs.”

Dig deeper: Composability has arrived, says MessageGears

The report paints a picture of martech where the only constant is change. This will be driven by evolving customer needs, but AI advancements should give marketers the tools to keep up with and sometimes get ahead of those changes. The full report can be downloaded here (registration required).

Finally, to give an idea of all the AI-powered changes in martech, Brinker and Riemersma did a genAI-only version of their famous martech landscape.