Claude-Powered Tier 1 Product Launch

This case study was created from an episode of the Product Marketing Adventure Podcast. It has been adapted into Dusted's case study format to preserve and showcase the guest's real-world experience. Written to preserve the voice of the true author, Ryan Radcliff.

When I joined Vivun, the company had deep roots with sales engineers—a full product line and customer base of technical people supporting big deals. The founders got fired up about a completely different problem: giving the seller expert support live on the call. Hero is our AI sales teammate that joins live calls, knows the deals, products, and customers, and helps proactively before, during, and after. Most sales tools work before or after the conversation; Hero is right there inside it. We burned the boats, built a whole new product for a completely different person and problem, and brought back the old Hero product name (we still held the copyright). This meant new branding, new colors, a new website, and a new way of thinking about Vivun as the overarching research firm on top of Hero. We’re around 75 people. My CMO sat down with the founders and E-suite first and locked the strategic narrative—an incredible gift—so I could start from there. I started about a week before Q1 on February 1st.

This was far more than a messaging refresh. It was a major identity shift for a new audience: quota-carrying sellers under pressure instead of sales engineers. We needed a whole new story, correct personas, and a new point of view, then had to operationalize a brand-new narrative from scratch and get a company of ~75 people flying in the same formation. The timeline was tight—soft launch April 6th, big launch May 5th, new website in two weeks. I also needed outputs that non-PMMs could run with without getting caught by hallucinations or scope creep, something my earlier ChatGPT attempts at a former company couldn’t deliver. Classic risks included assuming we already knew the persona or the industries where we win, plus the new urgency that LLMs would define us unless we showed up with extreme consistency.

I started with classic product marketing work, not AI: who it’s for, the problem we’re solving (aligned to how sellers see it), differentiated value, impact, proof, a positioning statement, a benefits ladder (emotional + functional), and a sanity-check filter—“if this doesn’t help a seller prepare, respond, or follow through when the pressure of the moment is on, it doesn’t belong.” I built a 40-page positioning and messaging manifesto plus a separate persona doc, doing heavy research with Claude as the foundation, then operationalized it across the company. The manifesto opened with purpose (“this is a positioning thesis”), then 12–13 sections of foundation: market context/why now (no patience on sales calls; the moment as the unit of value; cost of missing it; why existing approaches break as AI raises expectations), target personas, core situation (“when would I use this?”), old way/new way, what’s required to win, how Hero meets it, differentiated value, proof points, and key messaging. I front-loaded Raskin-style narrative so we defined market reality and the problem before ever talking about Hero or Vivun.

I created three core companion docs: the manifesto/messaging framework, a deep living persona research doc on the quota-carrying seller alone in the moment, and copy guidelines on how we write. Early testimonials plus GetWise interviews (questions like “what are the two things you have to do before a call?”, “what stresses you during a call?”, “walk me through the five minutes after”) and daily UserTesting surveys (25 sellers from around the world) fed the persona work. I threw transcripts and spreadsheets into Claude projects to synthesize the full picture and keep feeding market signal to product.

I loaded the manifesto, persona doc, copy guidelines, and executive narrative into a Claude project—essentially my PMM team and a small language model stuffed with Fletch PMM frameworks—so every output started from the same DNA. Feedback loops were constant: run new research against the messaging framework, critique the summary, tighten line items, and let the system push back when ideas drifted into generic time-savings claims. For buy-in I built a tight 9-slide executive deck (Claude drafted the first version): remind everyone we already aligned on strategic narrative, surface urgency (including how LLMs will define us), show the one-page framework, a before/after comparison table (Gong records the call; we’re in the call with you), how each team uses it, the shared litmus test, the contract for keeping the LLM gap closed all year, and how we’ll measure success. Once approved, we put the framework to work: Page frames → HTML dumped into Claude with the full context for “Mad Lib” homepage and persona pages (sellers, managers, Gong/Agentforce comparisons), booth messaging, blogs grounded in our research, analyst prep (Forrester-ready in 15–20 minutes), short character-count versions for Crunchbase/G2/AppExchange, and plain-page scripts for 1-, 3-, and 5-minute explainer videos. Creative leaps like billboard copy stayed human; long-form and structured assets ran through the Claude project with me as editor.

Claude was the core system—I treated project tabs as my PMM team and built a de-facto small language model with the 40-page manifesto, persona research, copy guidelines, executive narrative, and Fletch PMM frameworks so it could Mad-Lib HTML, draft blogs, comparison pages, video scripts, analyst briefs, and listings while holding the DNA and pushing back on drift. GetWhys ran targeted seller interviews; UserTesting delivered 25 survey responses a day from sellers worldwide. Page frames produced website structure I downloaded as HTML and fed to Claude. Google Drive held the living docs. I drew on April Dunford’s positioning sequence, Anthony Pieri’s “when would I use this?”, Raskin-style strategic narrative, and continuous customer conversation habits. The stack turned research transcripts and spreadsheets into synthesized persona insight and kept every artifact on-message.

The framework became the consistent DNA for everything. My growth-marketing partner used it as a primary input and shipped an entirely new website in two weeks. We soft-launched April 6th and did the big launch May 5th with the company aligned. Claude routinely delivered ~75% ready copy that needed only targeted tweaks, eliminating scope creep on pages that normally stall. We produced the three explainer videos in a week, spun out booth messaging, comparison pages, research-backed blogs, analyst prep, and all the directory listings without death-by-a-thousand-cuts. I could context-switch and 3x the work while the system kept claims accurate and on-strategy. The same living persona research now regularly surfaces industries where we win that we hadn’t fully seen, and I show up to product with real market signal. The executive contract—reinforce every quarter, measure, keep the LLM gap closed—turned a one-month reset into the operating system for the rest of the year.