AI can make agency work faster, but speed creates a new editorial problem: More output can also mean more mediocre output, a.k.a. AI slop. As AI-generated content finds its way deeper into everyday workflows, agencies have to think carefully about what happens to quality, judgment and institutional knowledge when the machine starts doing more of the work.
The real opportunity—and challenge—lies in building smarter guardrails around the technology without losing the human ingenuity and creativity that makes agency work so valuable. Here, members of Forbes Agency Council share how their agencies keep AI useful, accountable and firmly connected to the human judgment behind strong client work.
1. Original Creative Thinking Starts Before Using AI
Most teams use AI to combat writer’s block. Yet, by doing so, they outsource the hardest part of thinking. We encourage our teams to ban AI for first drafts. Staring at a blank page has the biggest value. It’s where original ideas emerge, perspectives take shape and your team’s creative muscles grow. We use AI later, though, to polish and scale what we have created. – Nataliya Andreychuk, Viseven
2. Sharper Prompts Produce Better Reading Value
We refuse to let AI spoon-feed us. To ensure a net gain, our agency completely avoids cliché prompts and invests significant time in sharpening them with firsthand data, raw outlines and strict audience constraints. We explicitly train LLMs to act as seasoned, niche-specific SEO experts. This rigorous process ensures every output is deeply valuable for human readers and free of AI slop. – Vin Sonpal, CS Web Solutions
3. Constant Workflow Improvement Facilitates Lead Gen
We’re all figuring this out constantly, and it’s really about the incentive. If you’re incentivized to create a lot of content, then it’s easier for people to dismiss slop. Since we’re incentivized to help companies with lead generation, we know the quality of the content has to be high. This forces us to be intellectually honest and keep improving a workflow until it reaches the quality we need. – Araminta Robertson, Mint Studios
Forbes Agency Council is an invitation-only community for executives in successful public relations, media strategy, creative and advertising agencies. Do I qualify?
4. Clear Rules Establish Personal Accountability
We established a widely communicated and reinforced “AI Manifesto.” This sets out our ambitions for AI, what constitutes good use and personal accountabilities for quality output. – Tim Wragg, Human8
5. Higher Standards Turn Efficiency Into Excellence
Ask a simple question: Is AI making you more efficient, or more effective? Through training and education, we push our teams to strive for new ways of using AI that holistically improve the quality of their work. Anyone can use these tools to produce more, but we believe they should be used to raise the bar. If craft and creativity are unique differentiators, then companies can’t let AI slop dilute their impact. Efficiency may be a by-product of AI, but excellence is our goal. – Dani Mariano, Razorfish
6. Senior Review Protects Client-Facing Work
AI is a tool used daily at our agency for research, competitive scans, GEO analysis and citation tracking. One nonnegotiable is that no client-facing work ships without a senior practitioner owning it. AI drafts; humans judge. We have primary sources for every statistic and voice audits before hitting send. – Ronn Torossian, 5W Public Relations
7. A Shared SOP Prevents Sloppy Output
You can’t monitor every piece of copy that goes out, so we don’t count on catching it. We built a detailed skill file, our standards, do’s and don’ts and quality bar, and every person and AI tool works from that same standard operating procedure. Start everyone from one defined standard, and most slop never gets created. Then a human does the final pass, always. Someone has to own the work, especially for a client. – Vinny La Barbera, imFORZA
8. Fresh Ideas Lead The Content Creation Process
Use an “AI last” methodology for new content. Start with your own, fresh ideas rather than querying generative AI. Then, use it to augment your ideas and drill down several layers with probing questions and instructions. Finally, validate all information is correct rather than hallucinated. – Dennis Consorte, Consorte Marketing
9. Human Judgment Still Guides The Work
I often say, “Tools are tools; just don’t be one.” AI is simply another tool. In the end, people need to take responsibility for what comes out. If you spend more time fixing AI-generated content than you save, it is not an efficient tool. I mostly use AI for research and brainstorming, but strategy, judgment and final approval always end with humans. – Mike Demopoulos, hosting.com
10. Better Decisions Count More Than Saved Hours
We don’t measure AI by hours saved. That’s a trap. We measure it by decisions improved. AI handles the blank page; our team adds strategy, context, design and taste. Cheap, fast content is easy to make and even easier to spot. Creative critical thinking drives real outcomes. In the end, tracked results, not efficiency, tell us whether AI was used well. – Monica Alvarez-Mitchell, Pulse Creative, LLC
11. Thoughtful Briefs Shape Stronger Results
The safeguard is in place before the model is ever used. A strong brief and a clear outline determine the quality of the output far more than the tool itself. When the input is thoughtful, AI becomes a genuine starting point our team can refine, rather than a shortcut that quietly lowers the bar. – Amy Packard Berry, Sparkpr
12. Deep Context Gives AI Content More Meaning
I encourage my team to use AI to improve on their context and key points—at the moment, I see a lot of AI content which, when you push on it, means nothing. The best stuff from AI has deep context and specified outputs, but it is easy to just lean on it. Retrospectively, I look back at some of the AI work we did a few years ago and cringe at how generic and flat it feels. – Dean Seddon, Maverrik
13. AI Accelerates Insightfulness With Human Direction
We treat AI as a tool for acceleration, not authority. The measure is simple: If AI makes the work faster but less insightful, it has failed. Every output must be reviewed, refined and improved by a human before it reaches a client. AI should remove busywork, while strategy, judgment and original thinking remain firmly human. – Simon C. Lee, Verta Ventures
14. Visible Accountability Protects Brand Trust
We keep AI useful by making it earn its place in the process. Every team member must show what part came from AI, what they changed and why the final work is stronger. That audit trail exposes lazy thinking fast. If AI saves time but weakens judgment, voice or trust, we pull it back. Efficiency matters, but reputation is the KPI that decides whether or not the tool stays. – Vaibhav Kakkar, Digital Web Solutions
15. Human Ownership Keeps Drafts Client-Ready
Our simple rule: AI drafts; humans own. Nothing ships to a client without a person editing it and putting their name on it. We use AI for speed on first drafts, research and repetitive work, then apply real data, client voice and judgment on top. If it reads as generic, it gets rewritten. AI is a net gain when it saves hours, not when it replaces thinking. – Nicholas Cormier, Home Builder Marketers
16. Editing Time Reveals False Efficiency
We measure it simply: Does the AI output need more cleanup than it saved? If “yes,” we drop it for that task. AI writes drafts, a human owns the publish button, and no one publishes what they cannot defend line by line. Net gain comes from pointing AI toward where judgment is cheap and humans where judgment is expensive. The moment editing AI takes longer than doing it yourself, the efficiency was fake. – Tessar Napitupulu, Arfadia
17. Clear Policies Keep AI Use Disciplined
This is a critical effort, especially with increased criticism and skepticism around the use of AI. We have an AI policy and ensure all team members adhere to it, and that it’s updated as the technology evolves. We share internally how we’re leveraging AI for optimization so everyone can grow and leverage it. We set boundaries as it relates to creative and writing, leveraging it for idea expansion, summarization and so forth versus the first step in creation. – Samantha Scott, Pushing the Envelope, Inc.
18. Accountable Editors Protect Published Work
The rule at our agency is that AI drafts, and humans decide. Nothing ships without a person who knows the account editing it and every stat being verified against the original source, because AI will confidently invent numbers. We also train the team to use AI for research and first drafts, not final output. The efficiency gain is real, but only if a human is accountable for what publishes. – Tony Adam, Visible Factors
