As marketing teams head into Q4 2026, the question is no longer whether AI has arrived. The more useful question, according to the latest MarTech analysis, is whether it has actually changed how marketing work gets done.
Adoption is nearly universal, but workflows are stuck
More than 90% of marketers now say they use AI, and most have an LLM of choice. Content creation remains the most common use case, with teams also building apps, personalizing experiences and enhancing workflows.
Yet the analysis points to a persistent chasm: moving from individual LLM prompts to AI that lives inside workflows. Many marketers are still literally cutting and pasting from an LLM because connecting that tool to a CRM, MAP or productivity suite is technically challenging. Integrations require IT approval, technical resources and teammates fluent enough to use what gets built. Cutting and pasting is easy, but it is not a workflow.
Smarter silos, not connected systems
Almost every marketing tool now ships with some form of AI, from elegant assistants to bolted-on features. That includes everyday tools like Excel, Google Sheets and PowerPoint. But AI inside each app does not automatically make the apps work together.
Instead of using AI to connect the CRM, marketing automation platform and productivity tools, many teams are creating AI-powered silos. For vendors, that is an expensive risk: they are paying to embed AI across their products, but marketers can only meaningfully use a fraction of those tools.
Data distrust and the Salesforce example
A deeper blocker is data confidence. Many marketers don’t trust the quality of their own data, and they know AI can amplify bad data into wrong customer messages. The analysis points to Salesforce’s Data 360 layer beneath Agentforce and its CRM as an example of getting this right. Data management and governance are expensive and can feel like scope creep, but bad data can sink AI initiatives. Garbage in, garbage out still applies.
More output, more rework
AI makes it easy to produce more content, designs, drafts and emails. But several marketing leaders caution that marketers are overfocused on volume. Customers don’t necessarily want more; they often want relevance, empathy and a human touch. On the content side, many marketers also report that checking AI’s prolific output creates extra work, not the promised efficiency. AI-generated work isn’t automatically efficient work.
ROI confidence is slipping
Proving brand ROI has always been hard. Proving AI ROI adds another layer, and fewer marketers are confident they can do it in 2026 than in 2025. The analysis warns that too many teams measure outputs — pieces of content created, drafts produced — rather than rework, integration costs, customer outcomes and business impact.
What to pressure-test next
- Workflow over tool: Ask whether AI is embedded in a repeatable process or sits outside it.
- Data before models: Fix data quality and governance before scaling AI output.
- Customer value over volume: Are you creating more because customers want it, or because AI makes it easy?
- ROI beyond output: Track rework, integration costs and customer outcomes, not just assets produced.
The final word from the analysis is clear: AI adoption has outpaced organizational strategy. Closing the gap is a leadership and data problem as much as a marketing one — and the old battles of quality vs. quantity, workflow efficiency, customer experience and business value haven’t gone away just because the tool changed.
Source: MarTech
