Product GTM · 18 launches in 21 months. Cadence, not project.

One PMM seat shipped 18 launches in 21 months at Tier 1 or Tier 2 impact.
Absorbed an SMS Inbox launch mid-flight after the GTM owner walked two weeks before ship.
Landed AI SDR at #18 Product of the Day on Product Hunt.
The monthly report tempo never broke, which is the part nobody hires for and the part everything else rests on.

For the voice-AI or agent founder reading this

Five of the 18 launches were voice or agent products: AI Voice Agent (Nov 2024), AIVA Actions (Jul 2025), AI SDR (Sep 2025, #18 Product of the Day), Outbound AI Voice Agent (Jan 2026), and a Holiday Agent in prep. The pattern transfers. The numbers are JustCall's, the shape is portable.

Launch the agent against the highest-anxiety buyer concern (compliance, hallucination, will it sound human) before the highest-novelty feature. Position the workflow it removes, not the AI it adds. Put the legal disclosure into the hero copy so the buyer's first read is consent, not magic. Track adoption per workflow, not per seat.

What the launches looked like

Two product pages the playbook shipped on.

What it was

One PMM seat against two Tier 1 engines (Sales Dialer and AIVA), with marketing, sales, and CS calibrating against the monthly report tempo. The brief: stop letting launches be events. Make them a cadence the rest of the org could plan around.

The diff

Before: launches shipped when the product team was ready, produced a hellobar and a blog post, then evaporated. No flight pattern. No downstream CRO read three months later.

After: a one-month flight pattern. MRD, LP, one-pager in month one. Launch assets and enablement in month two. A measured CRO or revenue read the quarter after. Sales Dialer relaunch (Sep-Oct 2024), AIVA Actions (Jul 2025), AI SDR Product Hunt (Oct-Nov 2025) all followed it.

What I actually did

1. Sales Dialer relaunch, Sep–Oct 2024. MRD, LP, video, blog, emails, pricing copy on web and in-app, UX copy for dialer modes, two enablement sessions. Downstream A/B in Dec 2024: Auto Dialer +44% leads, Power Dialer +55%, Predictive +28% (self-reported, monthly PMM report).
2. AIVA Actions, Tier 1, Jul 2025. Launch email cadence pulled a 51% open rate against a 30% newsletter baseline. Pricing change communicated same cycle. Sales and CS enablement same month.
3. AIVA Customer Webinar series fed the Actions launch: 200+ registrations, 125 attendees, 29 qualified leads, 24 qualified deals, 4 won, $2K MRR plus a $1K implementation package. First AIVA deliverable with revenue attribution.
4. Other 2025 lifts on the same report tempo. Workflows GA (160+ beta customers, ~400 workflows, ~25,000 calls). SMS Analytics launch email 45.2% open, 4.4% CR. SMS Bundles $67K+ MRR Jun–Sep. Build with AI LP 175+ submissions. AI SDR Product Hunt hit #18 Product of the Day.
5. SMS Inbox, Mar 2026. The GTM owner left. Absorbed the launch. Shipped 4 preview emails, 3 in-app banners, a What's New blog, 4 LP refresh drafts, 11 help articles, a CS FAQ, and rewrote the LP hero away from the "automation created the problem" frame into a buyer-first one. 200+ preview opt-ins before formal release.

What stayed honest

Every metric here is self-reported and lives in a monthly PMM report, available on request under hiring NDA. The losses sit on the same record: launches that didn't move a number aren't dressed up in the diff card, and comms work that produced nothing measurable isn't smuggled into the volume count. The Mar 2026 owner-departure absorption is the stress test. The cadence held. That's the part that isn't a metric.

What it became, in substrate

The launch flight pattern (MRD/LP/one-pager month one, launch and enablement month two, CRO read the quarter after), the monthly PMM report used as a planning instrument rather than a recap, the absorb-without-breaking-cadence protocol for owner-departure on in-flight launches, and the preview-announce-follow-up launch email cadence calibrated against the newsletter baseline are four of the moves substrate carries forward. The skill, routine, and principle layers are open source for any client to clone and run.