AI Tool Explorer: Peer Session with Lance Spence · PMM Jam 2026

PMM Jam 2026 / Mentors: Product Marketing POVs

Peer-Led Session

AI Tool Explorer: What PMMs Need to Know Right Now

A peer-led AI exploration with Lance Spence, frameworks, hacks, hard lessons, and the one principle that keeps you indispensable as AI reshapes the PMM role.

Lance Spence
Lance Spence
TV News Producer turned Product Marketer LinkedIn
PMM Jam 2026 Beta · Super Jammer
Session type
Peer-Led AI Exploration
Open discussion format with Jammers during PMM Jam 2026 Beta. No slides, no script, just real talk.
Topic area
AI for Product Marketers
Prompting frameworks, practical hacks, platform risk, and staying irreplaceable
Facilitated by
Lance Spence
PMM Jam Super Jammer, TV news producer turned PMM, and member of the Impact Award-winning Team EnCompass
What Jammers walked away with
A practical AI prompting framework, actionable hacks for better outputs, and a sharper perspective on owning your work in an AI-assisted world

The conversation.

This was not a presentation. It was an open discussion among PMM Jammers who showed up with real questions, real frustrations, and real experiences using AI in their day-to-day work. Lance kicked things off with a practical framework and the core principle behind it, then opened the floor. What followed was one of the more honest conversations of the week about where AI actually helps, where it fails quietly, and what it means for the PMM role long-term.

The session covered a lot of ground quickly. Here is what was on the table.

What we talked about.

AI is lazy by design. If you give it one to two sentences, it will take the shortest path to a result. That result will be generic. Lance illustrated this with a robotics example: an AI taught to walk a four-legged robot started walking it sideways because nobody specified what "walking" should look like. The lesson transfers directly to prompting.
The AI Onboarding Prompting Framework. Lance shared a structured prompting approach built around seven elements: Goal, Who, What, Why, Structure, Examples, and the Think step. Full framework available as a PDF download below.
Treat AI like a new team member. You would not hand a new hire one sentence and expect a finished brief. The same applies to AI. Context, background, and examples are not optional, they are what produce expert-level output.
The think step and Claude's scratchpad. Asking AI to show its reasoning before producing a final answer dramatically reduces hallucinations. In Claude specifically, asking it to work through a problem step-by-step before concluding helps you spot errors, check citations, and understand where the output came from.
AI as its own critic. Sushmita Banda raised a technique that stopped the room: ask AI to examine its own output for errors. Ask it to be its own detective. It will often catch citation errors, inaccurate stats, and sourcing problems that would otherwise require manual verification.
Terminology mismatch. AI uses technically correct language that may not match how your audience actually talks. Eve Horne shared a real example: at Experian, the term "design your flow" was used for a campaign management tool. To a marketing audience, "design" means visual creative. AI will not catch that kind of disconnect, you have to.
Conservative AI vs human judgment. Gary Dietz shared a case where AI recommended waiting until Monday to follow up with a prospect. His instinct said respond today. He followed his instinct and won the deal. AI applies general best practices. You apply context.
Platform ownership risk. Eve shared the story of being blacklisted from Google and Facebook during a 2020 ad campaign due to an algorithm misread, losing all the money she had put into a Black Friday push. Separately, Sushmita flagged a real case where an agri-tech company lost all their Claude work when access was shut down overnight. The point: you do not own these platforms. Build in a way that does not leave all your IP inside a tool someone else controls.
You are the talent. AI is the multiplier. The session closed on this. AI can automate tasks. It cannot replicate the cross-functional judgment, the audience instinct, or the strategic connections that make a PMM valuable. Use it to multiply what you are already good at, not to replace the thinking.

What to take into your next prompt.

Hack 01
Specify the voice with a brand reference

Instead of saying "write in a professional tone," say "write clearly and technically, like Notion's documentation" or "professional and direct, like Stripe's developer content." Named references give AI a specific target and dramatically improve tone consistency.

Hack 02
Add the Think step at the end of every prompt

Ask AI to show its reasoning before producing the final output. In Claude, using the extended thinking or scratchpad approach surfaces how it arrived at conclusions, making it easier to spot hallucinated stats, misattributed citations, and shaky logic before they end up in your deliverable.

Hack 03
Ask AI to critique its own output

After getting a result, ask: "Where might this be inaccurate? What assumptions did you make that I should verify?" AI will flag its own weak spots more reliably than it will just hand you something solid unprompted.

Hack 04
Feed it your top-performing examples

Give AI examples of content that has already worked: a one-pager that drove sales conversations, a campaign that outperformed, a customer story that resonated. It learns the standard from what you show it, not from what you describe.

Hack 05
Iterate and build, do not start over

Treat the first output as a draft conversation, not a result. Push deeper: "This is too generic, go narrower on the enterprise use case." "The CTA sounds like every SaaS product, differentiate it." The second and third prompts are where the useful work happens.

Hack 06
Back up your AI work outside the platform

Do not leave frameworks, research, or strategic documents only inside Claude, ChatGPT, or any other tool. Export, save to Notion, or store locally. Platforms can shut down access. The work lives with you, not with them.

Insight
AI will not save you from terminology mismatch

AI speaks in technically correct language. If your audience uses different words for the same concept, and in B2B tech, they always do, AI will not catch that gap. You are still the one who knows that "data lake" means something different to your sales team than to your engineering team. That institutional vocabulary knowledge is yours, not the model's.

"You don't build a house on a tool. You build a house with a tool. Use AI to multiply what you are already good at. You are the talent."

Eve Horne, PMM Jam 2026 AI Peer Session

Take it with you.

Lance built a one-page AI Onboarding Prompting Framework specifically for product marketers. It walks through all seven elements of a high-context prompt, Goal, Who, What, Why, Structure, Examples, and the Think step, in a format you can use immediately.

Framework PDF
AI Onboarding Prompting Framework for Product Marketers
Download the framework

Part 2: The tools in practice.

Part 1 covers why context matters and how to prompt for expert-level output. Part 2 gets into the specific tools, Lovable, NotebookLM, Gemini, Canva AI, Claude Code, and more, with real use cases and what each one is actually good for.

Companion piece · Part 2
AI Tools for PMMs: What Actually Works and Why
Read part 2

From Super Jammer to Impact Award Winner.

Impact Award Winner

Lance Spence was one of PMM Jam 2026's Super Jammers, Jammers who went above and beyond during the sprint week. He led this peer AI session, contributed to community discussions throughout the event, and then joined Sushmita Banda and Donna Cantrill as part of Team EnCompass, which won the PMM Jam 2026 Impact Award for their work on the Civilian Compass mission.

Their submission covered five interconnected deliverables: a Clarity Framework, Persona Canvas, Messaging Map, Homepage Redesign, and Channel Strategy. It was one of the strongest team efforts of the sprint.

See Team EnCompass's work →
Lance Spence
Lance Spence
TV News Producer turned Product Marketer · PMM Jam 2026 Super Jammer

Lance Spence is a Television News Producer turned Product Marketer. Lance operates PMM like a newsroom, ensuring stakeholders are aligned, deadlines are met, and everything is in order across a go-to-market strategy. He brings a unique perspective to PMM teams not only because of his writing skills, but also because of expertise working as a creative, which ensures that sales enablement and other marketing materials are well communicated across the marketing organization. He is a passionate product marketing leader who loves to partner with others and bring the best of his teammates to create the best work possible.

Back to blog

Leave a comment

Please note, comments need to be approved before they are published.