How to use AI in marketing (without fearing it will replace you).

As a marketing firm that seeks to create opportunities from change, we’re always searching for new ways to serve our clients better. While the launch of ChatGPT widely expanded the conversation about generative AI, many at our agency have been using AI software to complement our daily work since well before 2022.

And we’re not alone in implementation.

“Seventy-three percent of U.S. companies have already adopted AI in at least some areas of their business.”

Rather than share our opinions on whether AI will ruin creativity, we thought we’d share how we use it and where we don’t.

How E&S teams use AI.

Despite some commonalities across the agency, each of our teams uses AI in unique ways..

Management and operations.

When it comes to legalities, accounting and compliance-heavy areas, we rely on real-world, vetted sources and partners — not AI. But even our spreadsheet-loving coworkers are experimenting with AI in other ways to help with:

  • Verification of job descriptions
  • Data analysis of business trends and insights
  • Completing repetitive or mindless tasks
  • First drafts of job posts, accounting whitepapers and internal business communications

Information technology.

Machine learning and AI tools are built into most of our IT software, helping us to:

  • Scan for viruses
  • Monitor traffic
  • Analyze trend data

Client services.

It’s probably no surprise that our relationship-focused client services team has been the slowest adopter of AI. However, they are exploring how automation can eliminate repetitive tasks or support data-heavy ones, including:

  • Creating a content or project plan draft
  • Note-taking and transcription
  • Resourcing analysis and planning
  • Proofreading and QA
  • Searching for efficiencies in our processes


Most media-buying platforms are heavily driven by AI and have been for a while now. Some core ways our media team continues to engage with and expand AI usage include:

  • Automating bidding in auctions
  • Testing creative to find the right combination of ads to serve
  • Getting answers to media questions
  • Learning about media trends affecting audiences and platforms
  • Discovering new ways to prompt AI for better results
  • Auto-captioning videos

Growth and strategy.

Currently, this team has found AI works best as a sounding board for research that helps spark ideas and ensure we’re not leaving any stones unturned. Successful uses include:

  • Sparking research inspiration
  • Analyzing or summarizing information
  • Identifying prospective clients
  • Generating ideas for initial personas and attributes
  • Drafting questionnaires and discovery session outlines


All creatives know how daunting a blank page can be. Our creative team often uses AI to metaphorically throw spaghetti at the wall and eliminate a lot of bad ideas quickly. In addition to ChatGPT, the creative team uses Grammarly, Google Gemini, Midjourney and built-in Adobe AI tools to:

  • Brainstorm ideas
  • Generate specific images for use in comp creative, like storyboards
  • Extend the background of an image
  • Check grammar and spelling
  • Summarize articles
  • Analyze copy for voice, tone and personality
  • Get keyword ideas
  • Create “messy” first drafts of presentation flows and article outlines

Some tasks just aren’t meant for AI.

Our clients work with us because of the partnership, innovation and creativity that only our team can deliver. Clients expect a high level of originality, as well as confidentiality.

That’s one reason why we never use generative AI for final creative or brand work.

Everything you feed into the algorithm can be served to others. So, we never input confidential client information or data that could violate our NDAs.

There’s also the issue of copyright. As it stands today, work generated by AI cannot be copyrighted. We don’t want to create work for our clients that competitors could use with a few edits.

But beyond the legal reasons for keeping AI out of our final work, there are also quality issues associated with AI-generated art being in its infancy. Of course, the quality of the output of AI is impacted by the quality of your prompts. But even with the best prompts, generative AI today doesn’t currently meet our core value of being really good. Most AI art — at this point in time — is average at best.

The other area where AI is just average has to do with its inputs. Because automated responses are based on the averages of what it has been fed, AI can’t tell us anything new or innovative. Until it can, AI is not a replacement for novel research and market studies.

Three beginner tips for using AI.

Whether you’re still unsure if AI is right for you or you’re beginning to experiment with tools, here’s some advice to help you get started.

  1. Research different tools and software.

    There are AI tools for everything. Editing photos. Writing blogs. Checking writing for AI use and plagiarism. Making art. Translating scripts. And so much more. But not every tool delivers the same level of quality. While ChatGPT is a top choice for many because it’s free, free comes with a few downsides. ChatGPT doesn’t cite its sources or check for plagiarism. If you plan to use AI for content creation, it’s worth the money to explore software that does both of these things and more.

  2. Become a pro at prompts.

    Input generic prompts, get generic results. What does that mean? Ask ChatGPT to “write a 500-word article about business strategy.” Then, change the prompt to “write a 500-word article about business strategies that can improve customer acquisition and decrease costs in 2024’s volatile economy” and see for yourself. Some marketers will tell you the secret to good results is to pay for a list of prompts. But the reality is, nothing can replace your own trial and error. Don’t be afraid to refine your prompts multiple times. If you’re not getting what you’re looking for, consider if you need to reword, include more background information or try a different tool.

  3. Experiment. A lot.

    Don’t try one tool or one prompt and give up. Like any new skill, getting good at AI takes practice. Use the free trial period to test drive many different AI tools. Read the latest articles. Just keep going. If you give up before really putting the time in, you could be stuck doing manual tasks while your competitors automate.

Technology is changing. Blink, and it will change again.

Technological advancements continue to happen at an unthinkably quick pace. By the time this article is published, we’ll probably be exploring new AI trends and opportunities. But that’s why we love what we do. We thrive on uncovering new solutions and insights that can help us solve problems for our clients.

Looking for a partner who can help you stay on trend and get results? Start the conversation to see how we support your marketing goals — with and without the use of AI.

Leanne Prewitt

President & Chief Executive Officer

Shaped by her background in creative direction, Leanne leads the agency’s culture and creative vision and also oversees the operations that allow a team of marketing, design and media specialists to create powerful and effective work for their client partners.

Leanne began her professional career in New York City working for some of the nation’s leading agencies. In 2016, after a five-month sabbatical around the world, she returned to her hometown and joined Ervin & Smith. Her global perspective and expanded professional experience influence the work she does today.