By Stacey Piper
Every Halloween the big box stores and specialty popups seem to up the game with bigger, more realistic monsters and scarier automated creatures. As I was putting out our ever-growing yard-ful of artificial horror this year, the White House issued an Executive Order with standards and best practices aimed at labeling and regulating Artificial Intelligence (AI), the scariest—and arguably the most exciting—thing in marketing these days. AI is becoming more lifelike and expansive every day, and the President even proposed watermarks on AI-generated content, presumably so we can tell the tricks from the treats.
Nonetheless, just like a spooky, haunted maze, we embrace AI even more than we run from it. According to a May 2023 Salesforce study, 26% of organizations are using AI for marketing and sales, while 22% are specifically using conversational AI or virtual assistants. Generative AI tools are transforming the way we approach our marketing campaigns, from streamlining content creation to enabling more personalized customer experiences, and there are more to come. The Infosys 2023 Generative AI Radar report projects that U.S. and Canadian firms will invest $5.6 billion in generative AI projects in the next year.
Understanding Generative AI & Its Benefits for Marketing
Generative AI (GenAI) acts as a bridge between the user and the digital landscape, crafting new data based on a data set provided by the user. In other words, given the right human input, GenAI can serve as your personal researcher, writing coach, or designer, helping you work faster (and smarter?). Among marketers, 51% are now using generative AI, and 75% are seeking ways to leverage it to automate repetitive tasks and even help develop content.
According to the Salesforce survey, the biggest driver for using AI is productivity. 71% say GenAI will eliminate busy work and allow workers to focus on more strategic work, saving more than five hours a week. Imagine what that amounts to over the course of a year. In marketing, GenAI can speed up routine work, freeing up time for more creative and higher value tasks. In the 2023 Salesforce survey of more than a thousand marketers worldwide, the top five uses for GenAI are:
- Creating basic content
- Writing copy
- Inspiring creative thinking
- Analyzing market data
- Generating asset images
While I don’t recommend GenAI for drafting complex, technical material, GenAI can give you a kick in the right direction if you’re plagued with writer’s block, for example, by assisting with outlining and ideating. It’s also a handy extraction tool for creating templated content from your original content, deriving right-sized excerpts for meta tags and ads or summaries for social media posts. Some tools can do the same for graphics and videos.
Leveraging Generative AI in Your Content Marketing
When the human context or touch is clearly missing, or long delayed, GenAI is exposed for what it really is—the ghost in the machine. The Salesforce Generative AI Snapshot Research Series asked desk workers to identify the key elements to successfully using it in their roles:
- Human oversight (60%)
- Enhanced security measures (59%)
- Trusted customer data (58%)
- Ethical use guidelines (58%)
In my September 2023 blog, I touched on the need for authenticity in our brand interactions. There are poor uses of GenAI that should probably be avoided, and then there are uses that could be significantly improved with training and proficiency.
- Creativity & Originality: Since AI is based on pre-existing data, it can become overused, for example, for company logos or headshots. I’m already noticing everyone wearing the same outfits in AI-generated portraits. GenAI can lead to formulaic content that may lack authenticity—and believe me—your customers and prospects WILL notice. If your brand depends on exclusivity, originality, and creativity (whose doesn’t?), you should think twice before adopting generative AI, as it can compromise your identity as innovative and trustworthy.
- Quality Control & Validation: Don’t use AI if it overcomplicates the human-driven approach. Fact-checking AI-generated content can become even more time-consuming and resource-intensive than doing the research first-hand. An AI “shortcut” may necessitate an all-new process of SME validation and copyediting, adding additional steps to a process that was already working.
- Privacy & Ethics: As the President’s order on “safe, secure, and trustworthy” AI establishes, we don’t know what we don’t know. Generative AI models are trained on large data sets of text and code, including improperly shared personal data, and it can be difficult for marketers to control the output of the automated models. Don’t make yourself part of the fake news conversation or inadvertently spread incorrect, misleading, offensive, biased, or discriminatory information. Then you’ll have an even bigger problem to solve.
Three Tips to Fine-Tune Your GenAI Outputs
- Use the paid, specialized version of AI tools.
Paid GenAI tools are more powerful and feature rich, offering multiple languages, integration with other tools, access to proprietary databases, and customer support. With free trials, one-time use, and paid subscription models, you can test them and see which ones best serve your needs. Look for ones that are using OpenAI’s GPT-4, the latest and fastest language model with a larger and more diverse data set than GPT-3. GPT-4 can generate realistic and coherent text in a variety of formats for marketers.
- Test and refine your prompts.
Remember, the quality of what you put into a GenAI tool directly impacts the quality of the output. Invest some time in learning how to use your tool of choice, then carefully craft your prompt and build on it, one sentence at a time, continuing to generate results with each change. Some tools provide a template with input fields for each prompt, including type of deliverable or task (blog post, social media post, etc.), voice or audience (i.e., B-to-B or C-suite), tone (formal, informal, even a name such as an actor or politician). Keep in mind that the same prompt will provide a different result in another tool, so it’s a good idea to save your prompts and test for the tool that generates the most relevant results. For instructions and sample prompts for proposal language, check out the Rogue GovCon GPT Prompt SmartCard.
- Build your own AI dataset.
The intellectual property rights around derivative works are still being litigated, so it’s a good idea to inform your legal counsel of the scope of your AI use for your own brand and for client work. Look for an AI tool that allows you to populate it with your own brand material, thereby building your own dataset and leveraging your own IP library to train your AI platform.
Finding Your Way Through the AI Maze
Generative AI is transforming the marketing landscape, but beware of the tradeoff of speed and efficiency for the formulaic. GenAI can’t replace the contextual validation and creative expertise of you and your team, who have firsthand knowledge of your industry and your clients’ needs. While you may save time drafting a social media post or creating a logo, be prepared to invest time—and brain power—in populating the tools with meaningful inputs and ensuring accuracy and authenticity aren’t compromised in the outputs. Marketers who stay ahead of the curve and adopt generative AI tools will be well-positioned to succeed in the future as long as they take the time to learn the tools and adopt them thoughtfully and selectively.