T-10 weeks to T+19 weeksAI Product Launch Marketing Plan: The 10-Week Calendar

By T-3 weeks the assets have to be shot and the distribution accounts have to be eligible, because both take longer than launch week does. Moving a launch date takes an email. Producing a demo recording that does not exist, and aging an account that has never posted, takes weeks.

The asset list Lumina locks at T-10 weeks is four groups. That list and its T-10 deadline are Lumina's own campaign standard, not a published benchmark. Saying so matters here, because the agency pages competing for this topic publish their own launch numbers as though they were industry figures.

Eligible, in practice, means three things at once. The account has been posting for at least 90 days, which is Lumina's own threshold and not a number any platform publishes. Most of what it posts is content it made or was licensed to cut. It has not been repurposed from a different niche in the last quarter. Lumina treats all three as pass or fail at T-3 weeks, because a partial pass distributes exactly as badly as a fail.

Shooting and cutting are two separate deadlines and this calendar never merges them. T-3 weeks is when the footage exists and the accounts clear. T-1 week is when every launch-week clip is finished, captioned and queued in the scheduler. In Lumina's own campaign experience, and not from any published source, teams that collapse the two into a single "assets ready" milestone tend to discover in the final week that they have raw footage and no cuts.

The two dates are April 30, 2026 and July 15, 2025.

Instagram stopped recommending accounts that mostly repost other people's content. PetaPixel's Kate Garibaldi reported Adam Mosseri's announcement on April 30, 2026.

That is Adam Mosseri, head of Instagram, and the full detail on the change sits in Instagram's unoriginal content policy.

YouTube's inauthentic content policy, renamed from "repetitious content" on July 15, 2025 and clarified at the same time to cover content that is repetitive or mass-produced, makes that category ineligible for monetization under the YouTube Partner Program. A further clarification took effect July 16, 2026. TechCrunch's Sarah Perez, reporting on July 20, 2026, wrote that the updated guidelines break inauthentic content into three categories: "generic, repetitive, or template-based content", then off-putting or distressing content, then content where AI personas discuss sensitive topics like finance, legal issues, healthcare and medical issues.

An account spun up at T-2 weeks and filled with reposted launch footage is the exact pattern both rules describe. This is not a stylistic preference about originality. It is a distribution gate that closes before your clips ever reach a feed.

In launch week, AI buyers find you in three places: an AI chatbot answer, a review-site listing, and a clip on someone else's feed. Only the third is one you control.

The shift toward AI chatbots as the first stop is recent and measured. G2's "The Answer Economy", published April 15, 2026, found 51 percent of B2B software buyers now start their research with an AI chatbot more often than with Google. That survey ran in March 2026 across 1,076 buyers, alongside 39 marketer interviews. The comparable figure was 29 percent in G2's predecessor report "AI Now Means 'Always Included'", published May 14, 2025 from 1,169 decision-makers surveyed that April. G2's follow-up, the "2026 Buyer Behavior Report: The Evaluation Maze", published July 22, 2026 from 1,038 decision-makers surveyed in June 2026 plus 55 interviews, found 82 percent had sourced software recommendations from an AI chatbot in the previous 24 months.

The review-site listing is the surface most launch plans skip, and it also feeds the chatbot answers. Chatbots answer software-category questions using material that sits on review sites, so a category page with no listing on it in launch week is an absence in two places at once. Claiming the profile is a T-3 weeks job.

The third surface is a clip on someone else's feed, and it is the only one a launch team can put on a calendar. A chatbot answer depends on what has already been written about the category. A clip posted by a creator account with existing reach lands on a date you choose, in front of an audience that was not searching for you. That is why this calendar is built around the clip set rather than around the announcement. Whether a chatbot names you in launch week is testable, and the method is set out in how to get cited by ChatGPT.

Those last two G2 numbers change what a launch is for. They do not argue for a louder launch. They say launch week is an entry point into an evaluation the vendor does not control, so the assets have to survive being found second, by someone who arrived looking for a competitor.

Product Hunt belongs in the plan and its numbers do not. Product Hunt publishes leaderboards, not launch benchmarks. The upvote thresholds and launch-day traffic figures in circulation trace back to third-party aggregators rather than to Product Hunt, which is the same provenance problem set out in why benchmark numbers disagree. Plan the day. Do not plan against a number nobody published.

High Alpha's 2025 SaaS Benchmarks Report, ninth edition, drawn from more than 800 respondents with no survey completion date stated, records that "events, conferences, and real-world interactions dominated nearly all ARR bands". That line sits under a chart of the top five reported channels, so it is not a ranking against every channel. It is still the closest published comparison to what a launch week is: a company's one self-created event.

Day 0 ships the announcement and the first clip wave. Days 1 to 3 ship the answers to whatever objections surfaced on day 0. Days 4 to 6 ship proof. Nothing is produced during launch week, because everything was cut and scheduled by T-1 week.

I am Rhys McKay, and I run Lumina's clipping campaigns for AI and SaaS brands. What breaks in launch week is almost never the creative. It is the account. A client arrives with a strong demo, a sharp claim and a distribution plan built on handles that were registered six weeks earlier, and the clips go out into an audience of nobody.

The creative can be fixed in an afternoon. Ninety days of posting history cannot. That is why account eligibility sits at T-3 weeks in this calendar rather than in the launch-week checklist. This is my own read from Lumina campaign work, not a published finding.

The planner below returns one of three verdicts under the label Readiness against Lumina's production standard. No eligible distribution account reads *not achievable*, whatever the notice, because posting history accrues in real time. An eligible account with milestones still open reads *tight*. An eligible account with every milestone true reads *on standard*.

You cannot judge a launch on pipeline inside launch month. ICONIQ's "The State of GTM in 2026" reports the average B2B sales cycle fell to 19 weeks in the second half of 2025, having shortened by around six weeks over the year. That survey was fielded in January 2026 across more than 150 B2B software companies, answered by CROs, heads of sales, CEOs and heads of revenue operations. ICONIQ publishes the cycle length. The rule Lumina draws from it, that launch pipeline gets read at T+19 weeks and not before, is Lumina's inference and not ICONIQ's recommendation.

Three signals are readable long before T+19 weeks. The first is qualified entry volume, meaning how many of the right people arrived at all. The second is evaluation depth, meaning whether those people went on to ask for something. The third is whether an AI chatbot names the product in answers where it was absent at T-4 weeks.

Every threshold in the launch scorecard is measured against a company's own pre-launch baseline, because launch week has no industry number to measure against.

This 10-week calendar is the one Lumina runs for AI and SaaS clients. If your launch date is set and your distribution accounts are not already posting, that is the problem to solve this week, before anything gets shot. What a managed program covers is set out on our AI and SaaS clipping campaign page, what it costs is on pricing, and what it has produced is in the case studies.

G2, "The Answer Economy," April 15, 2026. 1,076 buyers surveyed March 2026, plus 39 marketer interviews.

G2, "AI Now Means 'Always Included'," May 14, 2025. 1,169 decision-makers surveyed April 2025.

G2, "2026 Buyer Behavior Report: The Evaluation Maze," July 22, 2026. 1,038 decision-makers surveyed June 2026, plus 55 interviews.

ICONIQ, "The State of GTM in 2026." Survey fielded January 2026 across more than 150 B2B software companies.

High Alpha, "2025 SaaS Benchmarks Report," 9th edition. More than 800 respondents, no survey completion date stated.

PetaPixel, Kate Garibaldi, reporting Adam Mosseri's announcement, April 30, 2026.

YouTube Help, Google, "Inauthentic content" monetization policy. Effective July 15, 2025; clarification effective July 16, 2026.

TechCrunch, Sarah Perez, "YouTube updates its inauthentic content guidelines," July 20, 2026.