Signals tracked for OmniThink AI
The signal classes we monitor for OmniThink AI, ordered by how close each sits to the moment a brand actually changes how it does product work. Every account on the accounts tab came out of one of these classes, and the class each account came from is named there.
How we rank a signal
Two questions decide a signal's value for a product-AI seller: how close is it to a decision about how product work gets done, and can the account actually buy at OmniThink's stage. A CEO committing to a shorter development cycle on an earnings call is close to the decision. A job posting mentioning AI is far from it. A founder-led brand can say yes to a pilot in a week; an enterprise takes quarters, however good the evidence.
So the screen runs two axes at once: signal proximity and win proximity. The published set deliberately leads with accounts that score on both, and labels the enterprise accounts as the long game rather than mixing them in unmarked.
Closest to the decision
The class that leadsPublic, dated commitments about product-cycle speed, newness, or assortment shape. These are owned by a named executive and carry a deadline, which is what makes them windows rather than themes.
On-record product-cycle resets
A CEO or CFO committing in an earnings call, investor letter, or annual filing to faster development cycles, higher newness penetration, or fewer SKUs. The numbers and dates are in the record, and someone senior is accountable for hitting them.
From this screenlululemon: cycles cut toward 12-14 months, 35 percent new-style penetration by spring, AI-enabled planning. Under Armour: 25 percent SKU cut confirmed May 12, 2026, another 25 percent announced.
Category entries at growth brands
A brand launching its first collection in a new category is standing up trend, colorway, line-planning, and sourcing decisions it has never run before, at exactly the moment the team feels the gap. At founder-led size, this is the most winnable window that exists.
From this screenLDMA's Contour, a fourth category in under three years. Bala's first-ever apparel collection, June 2026.
New product leadership with a lifecycle mandate
A first or newly created Chief Product Officer, head of merchandising, or design lead whose mandate spans trend through sourcing. New owners reshape process and tooling in their first two or three quarters; after that the window narrows. Tracked with a decay clock, not as a permanent signal.
From this screenVuori's first Chief Product Officer, effective April 20, 2026, held as watch because no system evidence has appeared yet.
Capital and program events
Product work follows the moneyA documented commitment whose downstream is product-organization work. Slower than tier one, but dated and dollar-backed.
Funding rounds naming product expansion
A raise whose stated use of proceeds includes new categories or product-development hiring. The signal is not the money; it is the named intention plus the pressure to show category velocity to the new investors.
From this screenBlissClub's August 7, 2026 Series B, raised explicitly to expand categories and strengthen product development. True Classic's strategic investment for category growth.
Private-label and own-brand program launches
A retailer standing up or expanding an own-brand program is building trend-to-line capability under retail calendar pressure, usually with a lean team. This is the class closest to OmniThink's demonstrated private-label work, and it is monitored continuously even though no account in the current set came from it.
Disclosed AI programs, used in both directions
A brand disclosing AI in product or planning work proves budget and appetite, and simultaneously maps the incumbent. We use this class to qualify the entry angle and to disqualify accounts where the incumbent already owns the seam.
From this screenRalph Lauren's predictive-buying pilot expanding in fiscal 2026: appetite proven, incumbent present, held as watch with the angle narrowed to what the pilot does not cover.
Turnaround strategies naming trend innovation
A new CEO's published strategy that names trend innovation, assortment simplification, or newness as pillars. Slower than an earnings-call commitment because the organization is still forming, but the direction is public and owned.
From this screenBath & Body Works' Consumer First Formula: sharper trend innovation, simplified assortment, category exits.
Confirming only
Never lead with theseUseful to strengthen an account that already has a tier-one or tier-two record. Alone, each produces the generic outreach every product leader already deletes.
Job postings naming systems or AI
A posting that names a PLM, planning stack, or AI adoption confirms the seam and the systems in play. It does not originate a window: the hire may be the whole answer, and a posting says nothing about external vendor appetite. This class never leads a list; it strengthens one.
From this screenThe Merrell product-analytics role: live, newly created, names Centric and AI-assisted reporting. Screened and held as a confirming signal, with a validated group-president route banked for when a stronger window appears.
Trend features and collection press
Discovery inputs that point at operators worth screening. Every published account traces to a primary record, not to press about it.
Three rules the screen enforces
Why the list is shaped this wayThe compounding rule
One signal is a lead; two independent signals at the same account is a qualified window. lululemon carries an SEC filing, an executive commitment, and program coverage. A single job post carries nothing but itself.
Route banking
When a validated decision-maker route exists but no live window does, the account is held, not padded into the list with an invented story. Most of what the screen surfaces on any pass is held exactly this way: watch windows on a decay clock, banked routes, and contact gates that failed and killed otherwise good stories. An account gets published when its window is open, not to fill a page, and it re-enters the moment its window does.
Calibration by what converts
Signal classes are reweighted by OmniThink's actual outcomes: which classes produce replies, which produce pilots, which produce closed revenue. The tiers above are the starting configuration, not a fixed taxonomy.
What each account arrives with
- Company and the specific signal that surfaced it, linked to the primary source with its date
- Why it represents a product-workflow opportunity, and what the record does not prove
- Timing read: winnable now, watch with a decay clock, or long-cycle aspirational
- Decision maker with title, validated work email where one exists, LinkedIn, and an honest flag where validation failed
- The suggested first line and the correction, so outreach never claims more than the record shows
Contact records are delivered to OmniThink in full, validation method and date included. Nothing stays inside our system.
Signal classes describe categories of public record and the monitoring cadence we run. Per-account claims live on the accounts tab with their sources.