AI-Generated Content: How to successfully incorporate it into your marketing toolkit

Identifying the right use cases, and appropriate tools, along with knowledge of how to use AI and setting the right expectations from it are key to successfully leveraging AI in marketing

Contrary to the sense we get from our LinkedIn feeds filled with posts about ChatGPT, AI-generated content, and LLMs, the actual adoption of AI in regularly generating marketing content is still pretty low. According to “The state of AI in 2023: Generative AI’s breakout year”, published by McKinsey in Aug 2023, only 14% of the respondents in the survey said that their organizations regularly used generative AI in their marketing and sales functions. Moreover, the survey also found that only 8% of the respondents said they were using generative AI for “Personalized marketing”, right alongside the two other use cases which were “Crafting first drafts of text documents” (9%) and “Summarizing text documents” (8%). So we are yet to see the real wave of AI-generated content coming our way. Overall, it is safe to say that generative AI currently serves as an assistant and efficiency booster for a relatively small group of early adopters in marketing. Noted, however, that 14% penetration of a technology that’s so new in such a short period is still quite remarkable.

After having used several AI content generation tools and building one myself, here are some thoughts on how to successfully adopt AI for content generation.

1. Identifying the right use cases

Although Generative AI tools based on LLMs can generate almost anything you ask them to, their utility and quality of output vary a lot depending on every brand’s and/or marketing team’s unique situation and challenges. Multiple parameters determine the uniqueness of the use cases.

Subject Matter

While one may be able to generate a pretty engaging piece of content on a topic such as “Top 10 Things to Do in San Francisco”, the same may not hold for a subject area that is deeply technical or requires niche expertise and data.

Resources and Bandwidth

Doing more with less has been the name of the game, and an excruciatingly over-used phrase, in the last year or so. Generative AI couldn’t have taken off at a better time though. While companies are down-sizing their teams, marketing not being an exception, the use of AI as a copilot for smaller teams doing a lot more is a blessing.

Augmenting Capabilities

Writing is an art, and not everyone can do it exceptionally well. Generative AI can be a great asset for marketers and non-marketers whose expertise does not include writing for an audience. From a technical founder wanting to share their thoughts and expertise on social media to a Google Ads specialist needing a starting point for text ads for a hundred keywords, AI can significantly augment their expertise with much-needed writing assistance.

Scaling Content Creation

Not all types of content creation can be, or should be, scaled with AI. But there are use cases where it can help. For instance, it is very ineffective, and detrimental to the brand, to scale blog article creation or social media posts with AI without significant human oversight of the content. On the other side, generating ideas for blog articles, and creating outlines that can then be expanded into complete articles with the assistance of AI or purely with human writing are great use cases for scaling. Similarly, improving and rephrasing previously written content to make it more readable and impactful is a great use case if done with some level of human oversight.

Every marketing team, or solo marketer/founder will have a slightly different mix of these parameters that apply to them and it is critical to identify where and how AI can deliver the biggest impact.

2. Selecting the right tool

Depending on your use cases, it is important to select the best Generative AI tool(s) that will get the job done. Like any other SaaS tool category, Generative AI tools vary a lot in their capabilities to tackle different content-generation tasks. Some are great at long-form content creation, while others do social media pretty well. Some generate a pretty broad range of content moderately well but don’t specialize in any particular format.

Additionally, the emphasis on brand and message pull-through varies significantly between tools. Depending on the type of content you are looking to create, you may want to look into a tool’s ability to understand your unique brand, tone of voice, capabilities, and differentiation in order to incorporate them into the generated content.

Some desk research and review sites can provide a solid starting point for exploration.

3. Learning Prompt Engineering

Anyone wanting to use Generative AI should spend some time understanding the basics of prompt engineering and playing around with the tools to get a feel for what works and what doesn’t in getting the highest quality content possible. Even though Generative AI processes and understands natural language, the right prompt can make a big difference in the quality of the generated content.

4. Not trying to game the system

AI-generated content is such a recent phenomenon that it is impossible to predict how content distribution channels are going to treat it alongside human-written content as the technology and its adoption evolve. There are tons of tools right alongside content generation tools that analyze content and tell you whether it is AI-generated, human-written, or a mix. Their accuracy varies a lot and there are plenty of false positives and negatives in identifying AI-generated content. Regardless, I think these are very dangerous to incorporate into Gen AI workflows. Remember when folks who created mortgage-backed securities had access to the same credit rating tools that the rating agencies used to rate the securities? It led to the inclusion of just the right amount of subprime mortgages in the security so that it passed the threshold for not being rated as junk. We all know how well that ended.

So regardless of the use of AI, the focus should always be on the quality, helpfulness, and relevance of the content to the audience, not on trying to make AI-generated content sound more like human-written content on whatever parameters that the difference is currently determined by the available tools.

5. Setting the right expectations from AI

While generative AI appears like a Swiss knife for content creation, it is important to acknowledge that it is, in fact, exactly that - a Swiss knife. It is a tool you can carry around with you and tackle a variety of use cases fairly well. But anything substantial still requires you to pull out that power drill or a sledge hammer. Moreover, to deliver impact through AI-generated content, a human still needs to play a significant role in the workflow. As the capabilities of generative AI tools stand today, they are not quite ready to tackle marketing on Autopilot. But they can be damn good Copilots.