The pressure to deliver higher output with leaner teams is a constant hurdle in modern marketing operations. While AI agents can take on research, reporting, and first drafts, delegating the strategic context behind that work is far more complex. An agent cannot recover the reason your brand avoids certain language or recall a delicate customer history unless your team captures it.

Navigating this shift was the focus of “Meet your new teammate: Building marketing workflows around AI agents” at the November 2026 MarTech Conference. Industry leaders explored how SEO, MOPs, and marketing leaders can share workloads with AI while protecting the human judgment that drives real business value.

Stacey Ackerman, an independent consultant and fractional CMO, moderated the discussion with Ed Poppe, founder and fractional marketing leader at Poppe Marketing; Sean O’Connell, head of customer success at Screendragon; and Lisa Shaw, founder of Devon Point Group. 

An audience poll underscored how widespread this transition is: 51% of respondents reported their organizations are actively piloting AI, and none reported avoiding it. Here is the path forward for making those pilots sustainable.

Document the “why” behind your marketing decisions

When headcounts shrink and departed teammates are not replaced, remaining staff face intense delivery expectations. Beyond the immediate workload, this creates a quieter risk: institutional knowledge walks out the door.

Shaw uses AI agents to coordinate tasks and compile reporting data, freeing her attention for high-value strategic work. 

“What AI is doing is it’s freeing up my attention so that I can spend it on judgment and relationships,” she explained. 

To keep her agents aligned, Shaw documents the reasoning behind past decisions — capturing brand nuances, previous customer friction, and historical context — and feeds those lessons directly into her AI instructions. “I think the why is just as important as the what,” she noted.

Poppe connected this practice to a core marketing strength you already have: writing a strong creative brief. Defining deliverables, articulating brand guardrails, and marking checkpoints for review remain essential when briefing AI. Your strategic documentation is what turns a generic tool into a reliable teammate.

Fix the workflow before automating the task

Accelerating one isolated task can inadvertently push bottlenecks downstream to other teams. O’Connell emphasized the difference between the overarching workflow — the end-to-end path work travels — and the individual agents handling repeatable steps inside it.

Across organizations ranging from solo experimenters to enterprise teams with agents in production, the biggest challenge is connecting disconnected pilots into a cohesive system. Rather than starting with a shiny tool, O’Connell advises pinpointing the operational bottleneck first and defining what success looks like. “Start with the process, not the technology,” he said.

Ackerman reinforced this approach: layering AI onto a broken manual process only magnifies the friction. As you build better workflows, Poppe encourages bringing cross-functional partners along. Sharing a working pilot gives hesitant colleagues in sales, ops, or creative teams a concrete model to learn from, turning isolated experiments into shared wins.

Keep customer-facing judgment in human hands

Drawing a clear line between agent execution and human ownership protects both your brand reputation and your customer relationships. Shaw offered a practical rule of thumb for delegation: “The closer a task gets to the customer, the more it needs a human in the loop.”

Agents excel at pulling data, formatting assets, and assembling initial drafts. However, executive communications, sensitive renewals, and messages requiring genuine empathy belong with people. O’Connell agreed that final signoffs must stay with human leaders—agents can accelerate the prep work leading up to a decision, not own the outcome.

Even well-trained agents require steady guidance, as output can drift off-brand over time. Shaw compares AI to an overconfident intern whose polished-looking drafts still require scrutiny. Building in a critical editorial review ensures your content stays genuinely relevant and persuasive for your audience.

Build quality checks and governance into your system

Sustaining AI performance takes ongoing calibration, and this is doable with lightweight guardrails. O’Connell shared how teams can deploy an evaluation layer — where a secondary AI model scores agent outputs against set brand criteria before flagging items for human review. Monitoring performance over time also helps you spot when switching models could improve accuracy or lower costs.

Governance also means partnering with your team rather than locking tools down. During the audience Q&A, O’Connell noted that strict bans rarely stop employees from turning to personal AI accounts to meet deadlines. Providing approved, secure access empowers your team to innovate safely instead of working in the shadows.

Building resilient AI workflows comes down to focused experimentation and open collaboration: pick one real bottleneck, define clear boundaries for your agents, and share what works across your organization. As Poppe summarized, “So learn and share.”