Korpora · Sample reports
What you get
AI assistants are quietly becoming the universal first-stop layer for both agents fetching data on a user's behalf and humans asking what to use. The brand that establishes mindshare in the current training cycle compounds inbound for the next two or three. The brand that does not gets locked out by the competitor whose name the model already knows.
Korpora measures organic AI mindshare and the three input metrics that drive it (human channel velocity, ecosystem health, agent-readiness), then ranks the top three actions by lift potential (channel weight × current deficit × ship feasibility × time-to-corpus). Each action scoped to land before the next training corpus closes. Three published examples below.
Three live samples
The operator dashboard is the deliverable: one live, working surface per brand, not a static PDF. Three published examples below, across three verticals. The long-form technical write-up behind each measurement is archived as a reference and linked from inside the dashboard's own methodology footnotes, never as a separate front-door report.
Operator dashboard · On-demand staffing
WorkWhile
Cross-channel measurement of the on-demand staffing category, rendered as a working operator surface: per-finding action cards, severity tiers, lift-potential ordering, and copy-pasteable engineering prompts. Operators lead every human-discourse channel, yet WorkWhile holds single-digit organic AI mindshare. This is what the measurement looks like inside a customer's team rather than a one-shot PDF.
Open →
Operator dashboard · Luxury fashion ecommerce
SSENSE
Five-stream cross-channel measurement of luxury fashion ecommerce: AI-assistant recall battery, AI demand signal, competitive Meta ad inventory, and Instagram + TikTok organic engagement. Six ranked action cards with engineer prompts, Wilson 95% CIs throughout, and a brand-risk callout on every card. No luxury retailer is cited at the domain level, the structural opening the cards attack.
Open →
Operator dashboard · Sleep tech / DTC
Eight Sleep
Cross-channel measurement of the sleep-tech category. Eight Sleep leads organic AI recall and runs the most Meta ads in the cohort, yet AI assistants cite eightsleep.com zero times while citing its rivals. Ranked action cards close that paid-attention-without-earned-memory gap, each with a paste-ready engineering prompt.
Open →
What every dashboard covers
Every operator dashboard renders the same measurement underneath, the same seven areas, surfaced as cross-channel streams and ranked action cards. The long-form technical write-up of each lives in the Korpora archive and is linked from the dashboard's own methodology footnotes, never as a separate front-door deliverable.
- 1. Executive summary. Organic AI mindshare headline, leading rivals on the same measurement, the two highest-priority findings.
- 2. Agent-layer mindshare (organic vs evaluation-stage). Per-model and per-framing breakdown, Wilson 95% CIs, spread diagnostic between the two metrics.
- 3. Human channel velocity. Post-cutoff vs pre-cutoff publication rate per channel (HN, Reddit, X, GitHub, arXiv, Substack), with Poisson 95% CIs. The leading indicator for the next training cycle.
- 4. Per-channel comparison. Where the moat is and where the gap is, broken out by foundation channel and per-rival.
- 5. Per-rival capability map. What each top rival claims publicly, what they ship, named customers, the content surfaces they invest in.
- 6. Engineering-actionable fix list. Six to ten prioritized actions ranked by lift potential, each with a target metric, effort estimate, and time-to-implement.
- 7. Methodology and audit trail. Every number reconstructable from the JSON sidecar plus the public methodology page.
On the JSON sidecar
Every dashboard ships with artifact.jsonbehind it. Same data, machine-readable shape, designed for the customer's own agent or analytics pipeline to consume directly. The sidecar is the contract; the dashboard is the same numbers as a working surface. Schema in the methodology.
Get measured before the next training corpus closes
For crypto, devtool, and infrastructure teams with an active engineering team. Submit your brand and a company email. Limited number of free measurements per month while we focus on our first design partner. Dashboard delivered in a few business days.
Engagements payable in USD, USDC, or your project's native token.
Related: methodology · how we work · home