5 AI Expectations Marketing Leaders Need To Set In 2025

Read the full article here: https://www.forbes.com/councils/forbescommunicationscouncil/2025/09/24/5-ai-expectations-marketing-leaders-need-to-set-in-2025/

Shaun Walsh, CMO at Peak Nano, is a marketing leader with expertise in grid-enhancing technologies, fusion, nuclear, cybersecurity and AI.

One of the benefits of becoming the chief marketing officer at Peak Nano was that they had no marketing department, which provided the opportunity to build an AI-era team and strategy from the ground up. Without legacy processes or expectations, I could assemble the right team while implementing an AI-centric stack and setting clear expectations with leadership on what a modern marketing team and workflow look like, as well as what AI can (and can’t) deliver.

Despite agents and vibe coding, AI tasks still live in silos.

AI is still in its infancy. In most cases, it exists as a series of distinct task-oriented tools operating in silos, requiring human intervention and integration (“cutting and pasting” between platforms). Many new tools automate a few functions but don’t effectively pass results from one system to another.

Yes, AI agents are beginning to integrate workflows, and “vibe coding” is helping my team build our workflows, but the underlying data structures and API automation links between platforms don’t exist. You need a programmer or integration team to make things flow end-to-end.

AI’s infancy will pass faster than you think.

Love it or hate it, AI is here, and it will completely transform how marketing teams operate. Rather than fighting it, I believe the trick is to evolve your skill set to work in parallel with it. AI won’t replace you directly, but people who harness it might.

That said, we need to establish clear expectations regarding AI's capabilities, timelines and costs. Shifting to AI is not free. OpenAI is on pace to hit $12 billion in revenue this year, and the VCs investing in AI companies expect significant ROI.

Leaders should set five expectations around AI.

CEOs from Amazon, Google, Ford, Duolingo and many other companies are all-in on AI, having bought into scale and efficiency predictions. They aren’t wrong, but the “when” and “how” for AI’s full-scale, full-throttle impact may not be based on the capabilities of AI as we know it today.

We’ve seen over-weighted expectations for new technologies before. To quote Gartner’s Hype Cycle model, we’re living through the “Peak of Inflated Expectations.” After a few years of testing, failures, successes and learning, here are five expectations I’d set with executives around AI.

1. AI Isn’t Free

Implementing AI requires careful planning and tool selection, securely leveraging internal data and access to more backend systems than most people think. Prompt engineering and AI agents need access to your data to perform work specific to your business.

Securing your data (using public LLMs makes it public) and protecting your IP is vital. Each AI query also comes with a cost. Iterative testing and API costs add up quickly, and many AI tools are frontends that query major large language models (LLMs). They have to pay, so you will too. You will pay to store your data, find the talent to operate AI and integrate platforms, especially if you want to incorporate AI with legacy systems.

You can use AI to forecast costs, but your business leaders must know it's not free, and there may be a period of double cost to operate legacy systems while you transition.

2. AI Does Tasks; Workflows Deliver Results

AI excels at specific tasks, but its outputs are usually trapped in isolated tools. Integration gaps are the AI world’s “dirty secret.”

Tools like Zapier and OpenAI Agent SDK can help map the gaps between tools, but finding and closing them requires human intervention. Educating executives on the necessary level of work, skills and updates is an ongoing effort.

3. AI Does Content, Not Context

Today, AI is the king of first drafts, creating marketing and sales content faster than ever before. While this simple fact could justify team reductions, the humans on my team are still at the heart of our marketing. They have turned into messaging and AI resource experts, indispensable to educating GPTs on our brand’s positioning, messages, purpose and context.

Just as we previously hired search engine optimization (SEO) experts to drive visibility in search engine results pages (SERPs), we now need people who understand how to make GPTs, Claude Projects, Gemini and Perplexity Spaces deliver results reflective of our brand.

4. The Quality And Consistency Conundrum

Public AI LLMs struggle with consistent quality, tone and branding. They can generate drafts, but need human editing to ensure usable, on-brand results. The expression, “your mileage may vary,” is apropos for AI’s generative work.

In creative, starting with AI and editing takes longer. We still need a graphic artist to create branding baselines, colors and diagrams. AI produces quantity, not reliable quality, so human oversight remains essential.

5. Building An AI-Enabled Team

AI teams require people who are brave, bold and curious, with a strong survival instinct. Those who are happy with the status quo will likely not survive in the AI world. They must be willing to move at the speed of change: fast.

You and your team need to get comfortable with ongoing experiments and occasional failures. Change your collective mindset, redefine the value of their roles as managers, teachers and operators of AI. Many things that got us our jobs are not relevant in the post-AI world.

Set the itinerary for your marketing organization’s AI journey.

I have learned two big lessons from my AI journey. First, don’t let AI happen to you. Make AI work for you. Leaders can harness AI to create next-generation marketing teams, attracting the right people to lead the jobs of the future. You must train your team and invest in those who are ready to leverage and evolve with AI. Otherwise, your team members will become the "cost-saving" casualties of the CEO.

Second, leaders must take control of AI expectations so executives understand what’s hype, what’s real and AI’s costs and measurable benefits. Lay out the roadmap, then move fast to execute on the expectations you set.

At the end of the day, the future belongs to marketing leaders who embrace AI as a strategic but imperfect tool. To stay competitive, build skilled teams, thoughtfully redesign workflows and take bold risks to harness the possibilities of AI.

Read the full article here: https://www.forbes.com/councils/forbescommunicationscouncil/2025/09/24/5-ai-expectations-marketing-leaders-need-to-set-in-2025/