Most product marketers are using AI wrong. They’re using it like a better Google Doc. A faster copywriter. A summarization machine. That’s like using a smartphone exclusively as a calculator. I know because I did the same thing for almost a year.
I’m Shafiq Shivji. I’ve spent 15+ years in B2B product marketing. Auth0, mParticle, CloudBees. I’ve built positioning from scratch, run launches that moved markets, and sat in enough win/loss calls to fill a warehouse. I’ve seen the discipline evolve through every major shift: cloud-native, PLG, the rise of the buyer, the death of the funnel.
None of those shifts comes close to what’s happening right now.
The Moment That Changed How I Think About PMM
Six months ago, I started building AI workflows to monitor competitors. Pricing pages, changelogs, job postings. Not a dashboard. Not a one-time pull. A system that runs on a schedule, cross-references changes against our positioning, and surfaces a brief with recommended updates to our battlecards.
Within weeks, it surfaced a pricing restructure that our existing competitive intel process had missed. It flagged new job postings in a vertical we were about to target, suggesting a competitor was making the same bet.
AI was doing in hours what used to take my team weeks. And it was doing it continuously, not quarterly.
That’s not AI as a tool. That’s AI as a teammate. And it forced me to rethink what product marketing actually is when the tedious, time-consuming work that fills 60% of your calendar just... disappears.
The State of Play: PMM Is Having an Identity Crisis
Let’s be honest about where we are. The numbers tell one story. The reality tells another.
Gartner says 81% of marketing technology leaders are either piloting or have already implemented AI agents. That sounds like the revolution already happened. But dig deeper, and you find that most of those “implementations” are basic. Summarize this call. Draft this email. Generate three subject lines.
That’s generative AI. That’s the copilot era. And the copilot era is already over for anyone paying attention.
The next era is agentic. And it changes the job description of every PMM alive.
Here’s the difference. Generative AI waits for you to ask it something. An AI agent decides what to do, executes a plan, uses tools, checks its own work, and comes back with a result. It doesn’t need a prompt for every step. It operates.
Think about what product marketers actually spend their time on:
Monitoring competitors across dozens of sources
Synthesizing win/loss data into patterns
Updating battlecards and sales enablement materials
Researching market trends for positioning work
Building launch plans from a template (again)
Chasing down product teams for release details
Making presentation slides pretty (yuck!)
Every single one of those workflows can be partially or fully handled by an AI agent today. Not in theory. Right now. I know because I’m building them.
What “Agentic PMM” Actually Looks Like
Let me get specific. This isn’t a thought experiment. These are workflows I’ve designed, built, and started running.
Competitive Intelligence on Autopilot
Old way: Quarterly competitive review. Someone spends two weeks pulling together a deck. It’s outdated by the time it ships.
Agentic way: An agent monitors competitor websites, G2 reviews, job postings, press releases, and social mentions daily. It cross-references changes against your positioning and battlecards. When something material shifts, it flags it and drafts recommended updates. You review and approve. Time investment: 20 minutes a week instead of 40 hours a quarter.
Competitive tracking tools have existed for years. But the agentic layer is what turns monitoring into action. It's the difference between a security camera and a security guard.
Win/Loss Analysis That Actually Works
Old way: Conduct 10-15 interviews a quarter. Transcribe them. Find themes. Build a presentation. Share it once. Nobody reads it again.
Agentic way: Every sales call is automatically transcribed and analyzed. An agent identifies win/loss patterns across hundreds of conversations. It can spot deals mentioning a specific theme in discovery and close faster. It can flag that a competitor objection appears in a majority of recent losses. These insights surface weekly, not quarterly. And they feed directly into updated talk tracks and enablement materials.
The insight cycle drops from 90 days to 7.
Positioning That Stays Current
Old way: Build positioning in a workshop. Document it. Revisit it annually (if you’re disciplined). Watch it slowly drift from market reality.
Agentic way: An agent continuously analyzes your market. It tracks how competitors describe themselves, how analysts frame the category, what language buyers use in reviews and calls. When the market shifts, it flags the gap between your current positioning and market reality. It drafts alternatives. You choose.
This turns positioning from a periodic exercise into a living system. That’s not a nice-to-have. In a market where AI-native competitors can reposition in days, annual positioning reviews are a death sentence.
Sales Enablement That Writes Itself
Old way: PMM creates battlecards, one-pagers, talk tracks. Sales ignores half of them. The ones they use go stale within weeks.
Agentic way: Agents generate enablement materials dynamically based on current competitive data, recent win/loss patterns, and the specific deal context. Before a sales call, the rep gets a briefing that includes the prospect’s tech stack, relevant case studies, the competitor most likely in the deal, and updated objection responses. Not a generic deck. A customized brief.
Highspot reports that 82% of B2B sales teams using AI-driven enablement platforms saw increased rep productivity. But most of that is still human-triggered. The agentic shift is when the system knows what sales needs before sales asks.
Why Most AI-in-Marketing Content Misses the Point
I’ve read dozens of “AI for PMMs” articles, playbooks, and courses over the past year. Nearly all of them fall into one of three buckets:
The Vendor Pitch. “Our tool uses AI to do X.” Useful for evaluating software. Useless for understanding how the discipline changes.
The Analyst Overview. “Here are 10 trends in AI marketing.” Informative but detached. Written by people who observe marketing, not people who do it.
The Fear Piece. “AI will replace marketers!” or “AI won’t replace marketers!” Binary takes that miss the actual story.
The actual story is this: AI agents don’t replace product marketers. They replace the product marketing model that’s 70% data gathering and 30% strategy. What’s left is the 30%. And that 30% becomes your entire job.
The PMMs who thrive in this shift won’t be the ones who learn to write better prompts. They’ll be the ones who learn to design, manage, and orchestrate agent-based workflows. They’ll think like systems architects, not just storytellers.
That’s a fundamental identity shift. And almost nobody is talking about it with the specificity it deserves.
The Gap Nobody’s Filling
Here’s what I noticed when I started looking for people writing seriously about this intersection:
Tool vendors write about their tools. Consultants write about frameworks. Analysts write about trends. Training companies sell courses on prompting.
Nobody is writing from the practitioner’s seat. Nobody is saying, “I run AI agents for my actual PMM job, here’s what works, here’s what breaks, here’s what it means for how we organize teams and define the role.”
That’s the gap. And that’s what Agentic PMM exists to fill.
I’m not building this publication because I think AI in marketing is interesting. I’m building it because I believe product marketing is about to split into two tracks. PMMs who operate with agents. And PMMs who get outpaced by those who do.
What’s Coming (and What Matters)
Forrester reported that in 2025, 30% of all B2B buyers viewed GenAI tools as a meaningful interaction during their final purchase decision. More meaningful than talking to a product expert. Read that again.
Buyers are already using AI agents on their side of the table. They’re researching your product with AI. They’re comparing your positioning against competitors using AI. They’re making decisions faster because AI compresses their evaluation cycle.
If the buyer is agentic and the PMM isn’t, you lose. Not because your product is worse. Because your go-to-market is slower.
Gartner predicts that by 2028, 33% of enterprise software will include agentic AI, up from less than 1% in 2024. That’s not a gradual shift. That’s a phase change. Marketing organizations that aren’t redesigning their workflows around agent orchestration will find themselves running yesterday’s playbook in tomorrow’s market.
The Thesis
Here’s what I believe, and what this publication will explore:
Product marketing’s next evolution isn’t about using AI. It’s about becoming agentic.
That means PMMs who design systems, not just deliverables. Who build agent workflows that run competitive intel, win/loss, enablement, and positioning as continuous processes, not periodic projects. Who think about their work as orchestration, not execution.
The best PMMs have always been systems thinkers. They connect product to market to sales to customer. Agentic AI doesn’t replace that skill. It amplifies it beyond anything we’ve had before.
But only if you build for it. Intentionally. Starting now.
Agentic PMM is the publication for product marketers who want to lead this shift, not react to it. Every post will include real workflows, real results, and real talk about what works and what doesn’t. No vendor pitches. No analyst distance. Just a practitioner building in public.
If that sounds like what you need, subscribe. The shift is already underway. The question is whether you’re designing the system or stuck inside someone else’s.
love how this goes beyond the high-level, oftentimes esoteric perspective on agentic pmm and actually talks about the architecture: listening to real events (sales calls, competitive intel streams, and more), constantly translating that into strategy and positioning updates, and having that strategy hub actually propagating into the enablement frameworks, agents, and surfaces that touch the GTM team.