The Generative AI Playbook for Product Marketing

WRITTEN BY Lyle Burns, Fluvio Consultant

Generative AI, as defined by IBM, is deep-learning models and algorithms used to create new content, such as audio, video, code, or text based on the data it's training on.

While AI tooling is taking the world by storm, AI usage is still nascent in the field of product marketing. AI tools are underutilized by PMMs; with only 53.6% of product marketers indicating they’re using AI tools to enhance their product marketing strategies based on a survey conducted by the Product Marketing Alliance. While the value of product marketing is in the deeply human-centric and centered insights product marketers provide, AI can be used as a tool to enhance the productivity of product marketers.1 According to a study conducted by the Boston Consulting Group recently on the effects of AI on productivity, consultants leveraging AI (GPT-4 specifically) completed 12.2% more tasks while doing it 25.1% faster. They also produced over 40% higher quality results compared to those not using AI.2

With this in mind, it seems that it would only be a matter of time until there is greater adoption of AI tools by product marketers. Let’s take a look at how product marketers can begin leveraging AI tools now and in the future.

Current Use Cases

Data Analysis for Market Research and Customer Insights

Product marketers must stay on top of what is occurring in-market and are always looking for inputs to create better content, refine go-to-market strategies, improve messaging, and create Ideal Customer Profiles and personas. AI can be used to analyze survey responses, interview transcripts, or transcripts from sales calls and quickly get key insights and summaries. This application can extend to sentiment analysis on responses on review sites and social media mentions to gather information. This isn’t limited to just your company or product either, it can also be used for gathering competitive intelligence to better understand the market as a whole. In fact, the competitive intelligence platform and Fluvio partner, Klue, uses AI to review and summarize insights, and recommend competitive alerts.

When building buyer personas I often will round them out using information from job postings to understand key problems people in a role are expected to solve, key tasks they have to complete as part of their daily functions and more. In addition, I will go on Reddit or other industry-specific forums to gather first-hand information on the buying process, problems they’re facing, top-of-mind topics, and more. Using AI, I can quickly go through these postings and forums to obtain summaries, identify themes and patterns, and discern key takeaways. From there, I decide which information is most relevant to supplement the personas I am creating.

Learning how to leverage AI for data analysis and market research now can put product marketers ahead of the changing landscape. According to a Qualtrics study on How AI Will Reinvent Market Research, 48% of market researchers expect that AI will take over statistical analysis within five years. In addition, 51% of researchers expect AI to take over conducting brand awareness tracking, and 62%3 expect it to take over reading open-ended responses within five years.

Content Ideations and Creations

Content ideation and creation is another way product marketers can leverage AI. ChatGPT in particular is an effective tool to use for brainstorming content ideas based on what your target audience may be looking for, high-performing keywords, popular content in the market, or other parameters. In fact, according to a study conducted by The Wharton School of Business, Chat GPT is 40 times more productive than humans at the task of idea generation, and the quality is considered higher than that of an average quality of an idea generated by a human.4

Once the ideas are generated, AI can be used to quickly generate a text draft that product marketers can then refine. It is important to check the quality of content drafted by AI and to check to ensure that the output is not plagiarized or copyrighted. However, this shouldn’t necessarily dissuade product marketers from testing AI for content production as a research study by the National Science Foundation, showed that “the average time taken [to write content] decreased by 40% and output quality rose by 18%”5. Maintaining human oversight and providing task-specific data can ensure quality while reducing risks of AI such as biased content, compelling, but inaccurate content, or unsuitable content. Product marketers should know their product and audience better than AI and should respect the knowledge of their buyers, so taking a fully AI-generated piece of content in most cases wouldn’t be most practical or effective for PMMs.

Project Management

Go-to-market efforts require heavy project management and coordination between stakeholders across the organization. Many project management solutions are incorporating AI in order to streamline and minimize manual processes. This includes AI-driven functionality such as risk tracking, AI-generated sub-task creation, and AI-written emails and comments to stakeholders. These tools can help more effectively keep GTM execution on track, keep stakeholders informed on their tasks and timelines, and aid in document management allowing PMMs to focus their time on strategic GTM needs.

Additional Potential Use Cases

While currently, product marketers can use AI to enhance their productivity and augment their workflow rather than completely replace or automate tasks, as AI matures the use cases will expand. These use cases can include:

ICP and persona creation
Customer segmentation based on market and CRM data
Defining market opportunities and problems
Product use case identification
Message testing and validation
Assessing product usage to provide customized messaging and prescriptive recommendations to customers and the product team

Risks

AI is still an emerging technology and as such using it comes with real risks that users should understand. AI requires human review and oversight of its output. The quality of the outputs can be dependent on the quality of prompts provided, how much data is available for the model to draw on, and the depth of expertise that is needed. With this in mind, the more niche or specialized a concept and the more specific the data required, the more human augmentation may be required.

In addition, generative AI still suffers from the standard set of problems often associated with AI including, potential bias built into the model, algorithm, or training data, the quality of the training data, lack of transparency, and potentially inaccurate outputs. With these in mind, users should be taking precautions and consistently checking outputs from AI, not letting it auto-run.

An often overlooked risk with AI is data privacy and security. This is critically important if a marketer is planning to use AI to analyze customer data that may include personally identifiable information (PII) or sensitive internal data. In fact last year Samsung had to ban the use of generative AI for their employees due to sensitive company data being leaked to Chat GPT.6 Data privacy companies, like Fluvio partner, Liminal, help organizations take advantage of the power of AI while protecting their data privacy and limiting security risks.

Conclusion

Generative AI has real potential to support the work of product marketers. However, it needs to be monitored and augmented with the human touch of intelligence to bolster the depth and quality of outputs and review for quality assurance. To make the most of AI as a tool, product marketers should be testing it in safe environments, studying prompt development, problem formulation, and determining approaches and guidelines for how to best use it. Product marketers who learn and develop AI skills now while it matures will be in a position to succeed and drive adoption at their organizations while understanding how to use it most thoughtfully and effectively with more limited risk.