Every design ask from every meeting, gathered and reconciled against the live system. Not what was promised — what is actually running, measured in the database today.
All 38 August tickets arrived as mirrors from the old project tracker. The design module is fully built and nobody is originating work in it. Adoption ran 27 in April, then 1, 5, 18 — and now zero. Everything below is downstream of that one fact.
These get conflated in every conversation. They have different owners, different systems, and different states. Treating them as one thing is why the asks feel endless.
Game-day graphics, CG backgrounds, print, content scheduling, designer hours. Lives in the Service Dashboard. This is the big one — 20,506 tickets of history.
Built · Not adoptedPre-sale renderings and pitch collateral. Explicitly not services work — belongs near Opportunities in the CRM. Natalia compressed his hub to a saved view.
Parked by NataliaReplace the outsourced design vendor. Brief plus assets in, proof out. Needs to work by late October to leave testing room before busy season.
Not startedEvery row below is built, deployed and reachable right now. The counts are live from the operations database — they separate what people actually use from what merely exists.
Everything that helps a person do their existing job faster got adopted — hours tracking at 23,988 entries, content scheduling at 3,036. Everything that asks someone to change where they work did not: proofs at 55, templates at 1, ticket files at 9.
That is not a product-quality problem. It is a switching-cost problem, and the switching cost has one specific cause — see below.
Measured, not guessed. Three findings from the live system that explain the zero.
Status changes notify by chat message only, and silently skip anyone without a chat account linked. 189 active staff, 90 linked, 99 not. All 189 have an email address on file. Alexis's team went back to the old tracker because things were getting lost — they were: the notification had nowhere to go.
Briefs have grown longer every year: 441 characters average in 2019, 879 in 2026, the richest on record. The one-line tickets are the old tracker's backlink and nothing else. The detail still exists on the other side and fills in as the team migrates. There is a genuine 6% of vague briefs — that is a real cost, and it is an intake problem, not a data-loss problem.
Nine files across twenty thousand tickets. Account managers move assets to storage by hand and paste a path. Some clients still ship a physical hard drive. The ticket is a description of work that happens somewhere else — which is exactly why nobody defends it as the system of record.
In the order that makes each next thing cheaper. The first item unblocks the migration; the migration unblocks everything else.
Reuse the branded mail path assignment notices already use; keep chat for the 90 who have it. Additive, small, and the single thing standing between Alexis's team and leaving the old tracker. Joe is already paying out that contract and killing it after this year.
Block submission until the creative direction, board specs and assets are actually filled in. Kills the 6% vague-brief tail at source, which is what drives the revision rounds that make graphics unprofitable. Daniel asked for exactly this in April and it was parked — Charlie independently arrived at the same problem in August.
Alexis's four remaining wishlist items are the actual price of migration: duplicate a request, templates per recurring client, the venue-tree sidebar, and day-before reminders. Templates already exist and have one record — the feature shipped without the content.
Charlie's new one and the only item with money attached — he wants it billed per credit and called the revenue guaranteed year after year. Brief plus assets in, proof out, into the existing approval flow. The pieces are unusually ready: the proof server is connected, the approval workflow is live, and image generation is already wired into the assistant.
The client review flow is finished and has 55 shares to show for it. Nothing here needs building — it needs to become the default path out of a design ticket instead of an option beside it.
Queued behind design sign-off by Alexis's own sequencing. Print has 21 records and is a skeleton. Do not start these before 01 through 03 land.
Full wishlist parked by Natalia and compressed to a saved view. Do not reopen without her — this is a political gate, not a technical one.
The storage holding every proof and every finished asset is about 55 TB, currently under 10% free with busy season starting, and has no backup — the spend was declined. The same leadership wants the data centre gone entirely at a quarter million a year.
Anything built on top of that storage inherits that risk. Worth putting in writing before the AI designer work starts, not after.