C
CAGA (Context-Aware Governance Agent) A parallel domain reasoner in the OAI³ pipeline. Each CAGA reasons independently within one domain — Narrative, Alignment, Human Capacity, Technical, Risk, Financial, Opportunity, or Market — and produces a Signal Record with a confidence score attached. CAGAs do not synthesize across domains; cross-domain synthesis belongs to MIA. Running CAGAs in parallel rather than sequentially is what allows OAI³ to hold contradictions across domains visible rather than collapsing them prematurely.CCF (Context Conditioning Frame) The context configuration that shapes your entire diagnostic from the moment it is set. The CCF is defined once at the start of each analysis and threaded into every reasoning agent in the pipeline. It carries your product identity, per-domain scope settings, and output constraints. Because the CCF propagates to all downstream agents, the quality of your intake directly determines the quality of your output — there is no correcting a poorly framed CCF after generation begins.
Check-in A progress report or context update you submit to MIA during your CoTerminal 30-day execution workspace. Every check-in is classified into one of three severity tiers that determine how the system responds:
- Tier 1 — Minor Update: Routine progress, completed actions, or small observations. MIA logs the update and suggests a next action. No admin involvement.
- Tier 2 — Material Context Change: A significant operational change — such as a shift in team, ICP, or financial position. MIA flags the update as material; admin may be notified depending on your settings.
- Tier 3 — Verdict-Challenging: Evidence that directly contradicts a finding in your Decision Record. MIA flags it for mandatory admin review. The verdict remains locked until admin evaluates the evidence.
CoLinear CosentriQ’s human-in-the-loop AI adoption testing product. CoLinear produces a Score across four dimensions — Output Trust, Reasoning Clarity, Experienced Accuracy, and Improvement Signal — by running your model’s outputs through DollarFifteen contributor evaluation and OAI³ synthesis. Available in three membership tiers: Validate, Refine, and Scale. See CoLinear overview.
CoDomain CosentriQ’s AI market intelligence map. CoDomain publishes versioned frames showing where financial investment and practitioner sentiment are aligned, diverging, or absent across AI product categories and industries. Available at codomain.cosentriq.com/map. See CoDomain overview.
CoTerminal CosentriQ’s strategic decision intelligence product. CoTerminal produces a Locked Verdict and Living Action Plan for AI founders and product teams navigating complex product decisions. Currently available by invitation or application via pilot. See CoTerminal overview.
CoVariance CosentriQ’s continuous execution monitoring product. CoVariance monitors how your execution variables are moving relative to each other over time — flagging drift before it becomes a crisis. Currently on the roadmap.
Concordant A CoDomain frame direction. Concordant means financial investment and practitioner field sentiment are aligned: the market believes in this category and is actively spending on it. Concordant frames represent the clearest market signal CoDomain publishes.
Confidence Range The range — for example, 72–84% — that expresses how confident the OAI³ system is in a given verdict or Score. Always expressed as a range, never as a single point estimate. A narrow range reflects high-quality intake with strong evidence signal. A wide range reflects thinner evidence and honest uncertainty in the output. The “What Would Sharpen This Analysis” section of your CoTerminal diagnostic shows exactly what additional data would tighten the range on your next cycle.
D
DollarFifteen CosentriQ-owned contributor brand powering CoLinear’s human evaluation layer. DollarFifteen contributors are not generic data annotators — they function as a validation intelligence layer, evaluating model outputs on behalf of the real humans the model is supposed to serve. Contributor quality is continuously scored and maintained. MCGA’s inter-annotator agreement check runs on every sprint before MIA synthesizes a Score.E
Execution Workspace The 30-day bounded environment in CoTerminal that activates after you accept your verdict. Includes a Living Action Plan, Leverage Move tracking, MIA chat, and a Decision Log. At day 30, you can initiate a new diagnostic cycle, export your workspace, or upgrade to CoTerminal Scale for continued access. See Execution Workspace.Export Ingestion CoLinear’s Tier 1 model access method. Upload your model outputs as CSV logs, JSON traces, chat transcripts, or prompt/output pairs — no system integration required. Teams that aren’t ready to connect a webhook or API can begin evaluation immediately with Export Ingestion and graduate to deeper access tiers when ready.
Experienced Accuracy One of the four CoLinear Score dimensions. Measures whether contributors — acting as proxies for real end users — experienced the model’s outputs as accurate within their actual task context. Distinct from technical benchmark accuracy, which measures performance against a fixed test set.
F
Field Sentiment One of two signal types powering CoDomain frames. Field sentiment is drawn from practitioner statements in public channels — LinkedIn, Reddit, conference talks, published articles — captured on a rolling 90-day window. Paired with financial investment data to determine frame direction.H
Human Signal Sprint The core evaluation unit in CoLinear. A structured run that: generates evaluation scenarios via Inverse OAI³, collects contributor evaluations through DollarFifteen, applies MCGA inter-annotator agreement across contributor data, and synthesizes a CoLinear Score through MIA. Also referred to simply as a “sprint.” See the CoLinear sprint lifecycle for sprint statuses and timelines.I
IAA (Inter-Annotator Agreement) The quality check MCGA runs across contributor data before MIA synthesizes a CoLinear Score. IAA flags inconsistencies between contributor evaluations and resolves conflicts — ensuring the Score reflects genuine signal rather than contributor noise. Contributor data that doesn’t clear IAA thresholds is flagged and excluded from synthesis.Improvement Signal One of the four CoLinear Score dimensions. Measures whether contributors could identify specific, actionable ways the model’s outputs could be improved. A high Improvement Signal score indicates your model is close to what users need; a low score indicates the gap may be more fundamental.
Intake The structured questionnaire completed at the start of a CoTerminal or CoLinear engagement. The CoTerminal intake spans 75 questions across 16 sections. The intake feeds the CCF and propagates into every downstream reasoning agent — intake quality directly determines confidence range width. Incomplete or ambiguous responses are the most common source of wide confidence ranges and admin delays.
Inverse OAI³ OAI³ operating in reverse. Used by CoLinear. Where Forward OAI³ takes real-world signal and synthesizes judgment, Inverse OAI³ takes a model’s claimed behavior and generates synthetic evaluation scenarios designed to test whether that claim holds against real human evaluation. The scenarios Inverse OAI³ generates are what DollarFifteen contributors evaluate during a Human Signal Sprint.
L
Leverage Moves The three highest-leverage actions in your CoTerminal execution workspace, ranked by impact. Leverage Moves are part of the Living Action Plan layer and can be updated as your situation changes — they are not locked in the way the verdict is.Living Action Plan The adaptive action plan beneath your CoTerminal verdict. Unlike the verdict itself, the Living Action Plan updates based on check-ins, completed steps, and new context you provide during your execution workspace. It includes Leverage Moves, step tracking, and a decision log. The verdict is locked; the Living Action Plan is not.
Locked Verdict The strategic verdict produced by CoTerminal after generation. Locked means it does not change based on user preference, pushback, or argument — it provides a stable decision anchor. If you flag disagreement during the acceptance flow, MIA explains the reasoning in detail. If you still disagree after review, an admin evaluates your evidence and may uphold the verdict, add a clarification, or (rarely) approve a regeneration based on documented intake error. The Living Action Plan beneath the verdict remains adaptive; the verdict itself does not. See Verdict Types.
M
MCGA (Multi Contextual Governance Agent) The cross-check layer in OAI³. After all CAGAs produce their Signal Records, MCGA receives the full set and produces an integrity overlay — surfacing contradictions across domains, identifying reality-grounding failures, and generating an evidence confidence score. MCGA does not render a verdict; it prepares the signal for MIA. In CoLinear, MCGA also runs inter-annotator agreement across contributor evaluation data.MIA (Model Intelligence Agent) The synthesis layer in OAI³ and the only component that produces a verdict or score. MIA receives all CAGA Signal Records, the MCGA integrity overlay, and relevant market context before generating output. In CoTerminal, MIA produces the Locked Verdict and Living Action Plan. In CoLinear, MIA synthesizes the four-dimension Score. MIA is also available as a chat interface in your dashboard — you can ask it to explain any section of your diagnostic or Score.
Money Ahead A CoDomain frame direction. Money Ahead means financial investment is running ahead of practitioner field sentiment — capital is flowing into this category faster than practitioners are adopting or publicly endorsing it. Often signals an emerging category where investor conviction outpaces real-world validation.
Mouth Ahead A CoDomain frame direction. Mouth Ahead means practitioner enthusiasm exceeds financial investment — there is significant public discussion and endorsement but limited capital deployment. Often signals a category that practitioners believe in but that hasn’t yet attracted institutional spend.
O
OAI³ The proprietary reasoning framework underlying all CosentriQ products. Defines how intelligence is collected across domains (CAGAs), cross-checked for contradictions and gaps (MCGA), and synthesized into a final verdict or score (MIA). Operates in two modes: Forward OAI³, which takes real-world signal and synthesizes judgment; and Inverse OAI³, which takes a model’s claimed behavior and generates synthetic test scenarios. See the OAI³ Framework page for a full technical explanation.Output Trust One of the four CoLinear Score dimensions. Measures whether contributors — acting as proxies for real end users — trust the model’s outputs enough to act on them in their actual work context. Trust is distinct from accuracy: a technically accurate output can still produce low Output Trust if it lacks transparency or confidence calibration.
P
Pivot A CoTerminal verdict type. Pivot indicates the current direction is not viable, but the underlying capability or insight can be redirected toward a different problem, market, or approach. One of five possible verdicts: Refine / Pivot / Rebuild / Validate / Scale.R
Reasoning Clarity One of the four CoLinear Score dimensions. Measures whether contributors could follow and verify the model’s reasoning — not just whether the output was correct. Models with low Reasoning Clarity scores may produce right answers in ways users can’t interpret, audit, or trust at scale.Rebuild A CoTerminal verdict type. Rebuild indicates that the current product direction cannot be salvaged through refinement or adjustment — a fundamental reconstruction is required. One of five possible verdicts: Refine / Pivot / Rebuild / Validate / Scale.
Refine A CoTerminal verdict type and a CoLinear membership tier. As a verdict: Refine indicates the core direction is sound but specific elements need deliberate improvement before the next stage. As a membership tier: The mid-tier CoLinear membership, between Validate and Scale.
S
Scale A CoTerminal verdict type and a CoLinear membership tier. As a verdict: Scale indicates the current approach is validated and ready for deliberate expansion. As a membership tier: The highest CoLinear membership tier, providing the deepest sprint depth and contributor specialization.Score (CoLinear) The structured output of a Human Signal Sprint. Delivered across four dimensions: Output Trust, Reasoning Clarity, Experienced Accuracy, and Improvement Signal. Produced by MIA after MCGA runs inter-annotator agreement on contributor evaluation data. Accessible through your CoLinear membership dashboard with MIA chat for deeper exploration. See The Score.
Sprint See Human Signal Sprint.
Sprint Lifecycle Statuses The defined sequence of statuses a CoLinear Human Signal Sprint moves through from intake to completion:
intake_submitted → task_drafted → needs_revision → approved → live → collecting → analyzing → complete → closed
See Sprint Lifecycle for a full description of each status.
V
Validate A CoTerminal verdict type and a CoLinear membership tier. As a verdict: Validate indicates the direction requires real-world validation before commitment — the signal is not yet strong enough to justify full investment. As a membership tier: The entry-tier CoLinear membership, designed for teams running their first Human Signal Sprint.Verdict The strategic direction CoTerminal produces for your AI product. One of five types: Refine, Pivot, Rebuild, Validate, or Scale. Expressed with a confidence range. After generation, the verdict is locked and does not change based on preference or argument. See Locked Verdict and Living Action Plan.