83% say fully automating content creation would reduce most or all of their agency spend
NEW YORK CITY, NY / ACCESS Newswire / October 1, 2025 / As 95% of marketing leaders face growing demand for content, 60% report that they are spending less on agencies this year due to AI, a survey conducted by Typeface ahead of Advertising Week NYC found.
Typeface surveyed more than 200 marketing leaders (of seniority of at least vice president) and found that 83% believe fully automating content creation would reduce most to all of their agency spend.
However, most have yet to realize the full potential of AI. The majority of marketing leaders (82%) say their AI agents are stuck in the pilot phase, unable to scale to full deployment. This aligns with findings from a recent MIT study, which found that 95% of generative AI pilots fail.
While 53% have reorganized their team or updated roles and responsibilities as a result of AI, 61% say that their team primarily uses AI at the individual level (vs. systems and platforms that enable collaboration). This individualistic use of AI could be preventing the enterprise-wide collaboration required to recognize ROI from AI investments.
“Too many AI projects stall in the pilot phase because teams treat them like tools, not transformations,” said Jason Ing, CMO of Typeface. “The data shows it clearly. Most marketers are still using AI in pockets, at the individual level, instead of rethinking how teams and workflows need to evolve. Real ROI only comes when leaders focus on the people side of change, including building trust, new habits and shared systems that make AI scale across the enterprise.”
AI has promised faster execution, but marketing teams remain trapped in slow, complex workflows, requiring, on average, ten different people to be involved and six different vendors or tools. More than two-thirds of respondents (66%) said that launching a new multi-channel campaign takes three to four weeks, even though 85% said in an ideal world, it would only take one to two weeks. Marketers getting campaigns out faster often sacrifice campaign effectiveness by only personalizing for certain segments.
Marketing leaders who have been able to successfully deploy AI agents at scale reported the following factors were critical to success:
High-quality data resulted in useful output from the AI agent (72%)
The AI agent worked with our existing systems and tech stack (55%)
Governance controls that ensured brand safety and consistency (52%)
Executive sponsorship of the initiative or deployment (48%)
Those who were not able to successfully deploy piloted AI agents at scale reported the challenges included:
Concerns around compliance, legal or privacy issues (56%)
Lack of technical resources or IT support (53%)
Poor data quality resulted in the AI agent’s output not being useful (48%)
Governance concerns related to brand safety and/or consistency (48%)
Cultural resistance to using AI agents or change management challenges (48%)
The AI agent didn’t integrate with existing tech stack or processes (47%)
Lack of executive sponsorship (22%)
This research is part of The Typeface Signal Report – a new industry insights series that spotlights the signals redefining marketing in the age of AI. Designed for CMOs, brand leaders, and performance marketers, the report delivers data-driven perspectives to help leaders navigate transformation with clarity and confidence.
Learn more on the Typeface blog.
About Typeface
Typeface is the first Campaign Orchestration Platform powered by Agentic AI. From a single brief and brand guidelines, Typeface enables Fortune 500 enterprises to create personalized emails, ads, web pages, and videos – all generated, approved, and optimized in one secure workspace that connects seamlessly with existing martech tools like Salesforce, Microsoft, and Google. Backed by Lightspeed, GV, Salesforce Ventures, Madrona, Menlo, and M12, Typeface has been recognized by Gartner, TIME, LinkedIn, Fast Company, and Adweek as a leader in AI for marketing.
Contact
Courtney Brigham
[email protected]
SOURCE: Typeface
