AMA: Clari VP, Solutions Marketing, April Rassa on AI and Product Marketing

I probably see it in 5 key areas: 1. Flattened Differentiation I’ve seen PMMs over-rely on AI to write content. The output is clean, but it strips away nuance and voice. Everyone ends up sounding the same. When messaging is generic, it becomes harder to tell one brand from another which is the opposite of what good product marketing is supposed to achieve. 2. Shortcutting Discovery and Insight Some PMMs lean on AI summaries of customer calls, win/loss notes, or analyst reports and stop there. AI ...Read More

I like AI for copy and messaging when it is used as an assistant, not an author. It speeds up drafts, uncovers options, and helps test variations. It is not a substitute for judgment, context, or voice. Why I like it Fast first drafts. It gets you past the blank page so you can spend time on strategy and nuance. Rapid iteration. You can generate 10 headline directions or three persona variants in the time it used to take to write one. Pattern detection. It surfaces language customers actually us ...Read More

Market Research I use AI tools to compress the time from raw information to actionable insight. For example: Perplexity + Gemini: I’ll feed in a new Gartner or Forrester report and have them generate a structured digest of key themes, vendor mentions, and emerging buzzwords. Gemini tends to be stronger at pulling in live web data and giving me quick fact-based answers, while Perplexity is better at stitching together citations across sources. Claude: When I drop in a 50-page PDF analyst report, ...Read More

I think it really depends on the type of company. If you want to be a product marketer at an AI-native company, where AI is the product, you need a baseline technical understanding. You do not have to code, but you do need to know how LLMs work at a high level, what embeddings are, or why observability matters in an AI pipeline. That gives you credibility with technical buyers like CIOs, CTOs, or developers. Your role is to take technical differentiators such as data security, observability and ...Read More

Most people are still using AI for obvious things like drafting blogs or emails. That’s the low-hanging fruit. It saves time, but it doesn’t fundamentally change how a business operates. The bigger opportunity is when you use AI to augment workflows. For example, I'd like to use AI to take PRDs from product and turn them into a tiered launch bill of materials. It automatically suggests the GTM assets we’ll need based on the scope of the release whether that’s a simple FAQ, a blog post, or a full ...Read More

Saying “I use AI” does not stand out anymore. Everyone says that. The differentiator is showing how you’ve used it and what impact it had. For example: “I used Claude to cluster 500 survey responses into three customer themes in one afternoon, saving the team two weeks of manual coding.” “I built a Glean agent that critiques product messaging against our corporate narrative, so every draft is stronger before peer review.” Hiring managers want evidence you’ve applied AI in your workflow, not just ...Read More

There are a few examples I'll share related to actually building out agents and continuously fine-tuning them. My team is playing with some of these: 1. For Product Launches Build an agent that ingests PRDs, past launch plans, and your tiered launch framework. When a PMM uploads a new PRD, the agent automatically outputs a proposed launch bill of materials (FAQ, blog, enablement deck, battlecard updates). 2. For Messaging Development Create a “message map critique” agent (similar to the Glean on ...Read More

Well, we're guilty of this too but it's not slapping AI-powered or AI-driven in front of your Marketing. Customers care about care about outcomes: faster response times, higher accuracy, lower costs, or smoother workflows. Differentiation starts by mapping the technical advantage to a business result that matters to your buyer. Second, frame around trust and reliability. Trust is as big a differentiator as speed or price. Can customers rely on your product to be secure, compliant, and consistent ...Read More

I'm not sure if these are frameworks, but my four buckets for defining trust in AI products include: TransparencyCan customers see how the system works, what data it uses, and where its limits are?Examples: Model cards, documentation, explainability reports. Reliability and OutcomesDoes the AI perform consistently, and does it actually deliver the results customers expect?Examples: Accuracy benchmarks, latency guarantees, safeguards against model drift, business outcome case studies (e.g., pipel ...Read More

This is a fantastic question because it gets at the tension every PMM feels in AI right now which is that the product is changing weekly, but GTM motions and measurement cycles are built for quarters. Focus messaging on what does not change: Who the product serves (the persona) What job it helps them do better (the workflow) Why it matters (the business outcome) Considering building lighter, faster signals into the loop: Weekly field intel calls to capture customer and prospect reactions in near ...Read More

What cuts through is having a clear, opinionated stance. That could be about trust, workflow integration, data ownership, or vertical expertise. A POV forces you to say: we believe this is the right way to run your business, and it makes your narrative harder to copy. Pulling in proprietary signals like usage benchmarks, win/loss patterns, or voice-of-customer research grounds your messaging in something only you can say. Generative AI tends to flatten tone. Everyone starts sounding alike. A dis ...Read More