Ten deterministic signal families — now including per-competitor USPTO filing pace and disclosed-capital posture, alongside live year-over-year publication trends, industry-authorship share, and talent mobility — plus an Opus-powered strategist brief with a contrarian thesis and the falsifiers to stress-test it. On every Pro analysis.
Every PRO Insights tab is built from the same evidence set as the rest of the dashboard. The numbers come from a deterministic engine with no model in the loop; the reading of those numbers is a separate, clearly labeled layer.
Computed per analysis by an engine with no model in the loop — every signal shows its math, its confidence basis, its sources, and a caveat where the data is thin.
Hover any family for the shape of its math — representative samples, not live output.
new Choose an insights lens — TTO or Corporate R&D — to tune which families lead and how the strategist brief weighs its moves. The signal math never changes.
An interpretive read on the signals — clearly labeled as interpretation, never dressed up as computation.
What most analysts looking at the same evidence will conclude.
Where we differ and why — argued from the signals, not vibes.
Non-obvious plays, each with a first step.
Falsifiable indicators you can actually watch for.
The conditions that would retire each thesis — stress-test it, don’t just read it.
The trend family charts full-corpus publication counts per year (OpenAlex), and the industry family tracks the share of works with industry-affiliated authors. Where USPTO data supports it, the trend adds a filing-window check — publication acceleration weighed against title-phrase filing counts, complete filing years only. Honest complete-year math: the current partial year never inflates a trend.
No black-box “momentum score.” Each year-over-year figure is a division you can check, computed from counts you can click into.
Every trendline carries its confidence basis — sample size and time window — and a caveat where the corpus is thin. Filings math uses complete filing years only: the ~18-month application-publication lag keeps the newest filing years out of the YoY.
On top of the signals sits an Opus-powered strategist brief. It states what the field believes, argues where we differ, recommends moves with horizons — and lists the falsifiable conditions that would retire its own thesis.
Edge-AI payload processing is crowded: publication volume is compounding, three primes hold the visible IP, and the safe read is “fast follower, license in.”
The crowding is in the wrong layer. Filings cluster on hardware acceleration while the 2024–25 publication surge — and both academia→industry moves — sit in on-orbit model retraining, where the patent shelf is nearly bare.
Premium intelligence is only worth the premium if you can check it. These rules are structural — they hold on every card, every brief, every analysis.
Evidence-set signals cite the exact papers, patents, and awards they were computed from; the longitudinal families cite the live corpus query and the author records behind them. Either way — click through to the underlying data.
“47% YoY = (412 works in 2025 / 280 works in 2024) − 1” — the computation on the card, not just the conclusion.
Every signal states its confidence level with its basis — sample size and time window — right next to the claim.
Thin data gets a caveat on the card, and every strategist thesis ships with falsifiable indicators and a “what would change our mind” panel.
Every PRO Insights run — and every successful run of a watched paper or grant topic, scheduled or manual, hourly to monthly — records a dated observation set into the longitudinal trend store, at most once per topic per day. Watches on any plan accrue that record from their first run too; reading it back is PRO: from the second dated observation onward, the dashboard bakes a “Since the last observation” panel into the PRO Insights tab, and PRO owners’ weekly in-app digest carries the same dated drift line on each paper- or grant-topic watch — even in weeks where the watch surfaced nothing new.
| Observed | Works (2025) | Filings (2025) | Industry share | Change on the record |
|---|---|---|---|---|
| 2026-01-12 | 402 | 39 | 22.6% | first snapshot — baseline |
| 2026-04-03 | 408 | 40 | 23.0% | +6 works · +1 filing |
| 2026-07-04 | 412 | 41 | 23.3% | +4 works · industry share +0.3 pts |
The table is the store’s contents; the panel above is the shipped read-back — live in Pro dashboards from the second dated observation onward, baked at render time with both observation dates named. Note the 2025 filing count back-filling across snapshots (~18-month publication lag): exactly why the filings YoY math uses complete filing years only — the drift panel reports the back-fill honestly instead of selling it as a trend. Continuous monitoring of saved topics still runs via Watches.
PRO is our premium tier — the intelligence layer on top of a base plan. PRO adds all ten deterministic signal families, the Opus-powered strategist brief, and longitudinal drift, on every PRO analysis — priced at 2× your base tier.
2× your $1,500 TTO base — the PRO Insights layer added to a TTO plan (3-seat minimum). Ten deterministic signal families, the strategist brief, and the drift record on every PRO analysis.
Talk to sales2× your Corporate R&D base ($250/seat/mo ≈ $3,000/seat/yr) — the same PRO Insights layer on a Corporate R&D plan, alongside its service and onboarding commitments.
Talk to salesProposed pricing — the PRO rate is 2× your base tier and isn’t locked yet. Compare all plans →
Tell us what your team is scouting and we’ll walk you through a Pro analysis — signals, math, sources, and the brief that argues back.