> ## Documentation Index > Fetch the complete documentation index at: https://docs.cosentriq.com/llms.txt > Use this file to discover all available pages before exploring further. # CoDomain Market Signal Sources: Sentiment and Spend > CoDomain frames are built on two signal types — practitioner field sentiment and financial spend data — to surface each frame's directional call. Every frame in the CoDomain map is built on two distinct types of signal: what practitioners are actually saying in public channels, and where money is actually flowing. These aren't proxies for each other — they measure different things, they come from different sources, and they can point in opposite directions. The relationship between them is exactly what the map's directional call surfaces. Understanding how each signal type works helps you interpret a frame's direction with the right level of confidence. ## The Two Signal Types What practitioners are saying publicly about a category — captured from LinkedIn, Reddit, conference talks, newsletters, articles, and surveys on a rolling 90-day window. Where money is actually moving — sourced from analyst reports, financial disclosures, funding announcements, and procurement data, tiered by source reliability. ### Field Sentiment Field sentiment captures what practitioners — the people actually building, buying, and deploying AI products — are saying in public channels. The sources include: * LinkedIn posts and comment threads * Reddit discussions (relevant subreddits and community forums) * Conference talks and published session summaries * Industry newsletters and practitioner-authored articles * Surveys with disclosed methodology and sample sizes This signal runs on a **90-day rolling window**. The map reflects current market conditions, not historical consensus. A frame's sentiment profile can shift meaningfully between versions as adoption evolves, skepticism accumulates, or enthusiasm fades. **Minimum threshold:** A frame requires at least **20 qualifying practitioner statements** to publish a field sentiment signal. Frames that don't meet this threshold move to Insufficient Signal status regardless of what financial data shows. ### Financial Signal Financial signal tracks where capital and procurement spend are actually moving in a given AI category. Sources are organized into three tiers by reliability: The highest-reliability sources. These include SEC filings, publicly disclosed ARR figures, and announced funding rounds. Tier 1 sources reflect actual committed capital or disclosed revenue — not projections or estimates. Institutional research from analysts including Grand View Research, Mordor Intelligence, MarketsandMarkets, Menlo Ventures, and comparable firms. Tier 2 sources involve estimation methodology, so they're weighted below Tier 1 but still treated as defensible signal. Survey-based procurement data and secondary financial sources. Tier 3 is used for context and directional support but doesn't satisfy the minimum sourcing threshold on its own. **Minimum threshold:** Each subpattern requires at least **2 Tier 1 or Tier 2 sources** to publish a financial signal. Subpatterns that don't meet this threshold cannot produce a directional frame. ## Sentiment Clusters Field sentiment statements aren't just counted — they're categorized. Every practitioner statement collected for a frame is classified into one of three sentiment clusters: * **Enthusiastic** — The practitioner expresses clear positive signal: adoption, recommendation, observed ROI, active deployment * **Cautious** — The practitioner engages with the category but hedges: conditional adoption, unresolved concerns, wait-and-see positioning * **Skeptical** — The practitioner expresses doubt: failed deployments, overhype claims, reluctance to invest The distribution across these clusters shapes the frame's signal story. A frame with 40 statements that are 90% enthusiastic reads very differently from a frame with 40 statements split evenly across all three — even if both technically meet the minimum threshold. When you read the signal section of a frame, look at the cluster distribution, not just the statement count. ## The 90-Day Rolling Window The 90-day window is intentional. It makes the map a current-conditions instrument, not a historical archive. This means a few things in practice: * A frame's direction can shift between versions as new statements enter the window and older ones age out * Sentiment that was enthusiastic six months ago doesn't carry forward if practitioners have gone quiet or turned cautious * Emerging categories can move from Insufficient Signal to a directional call quickly if practitioner conversation spikes When you're reading a frame, check the version's data refresh date. The 90-day window runs backward from that date — that's the practitioner conversation the frame reflects. ## How Signal Becomes a Direction The directional call for each frame — Concordant, Money Ahead, Mouth Ahead, or Insufficient Signal — is computed from the relationship between the two signal types. It isn't generated by a model or written editorially. The directional call itself reflects whether financial investment and practitioner sentiment are aligned, diverging, or absent. The CosentriQ team performs a quality review before each version is published to ensure every frame that reaches you is defensible. Field sentiment statements are collected, filtered for relevance, and classified into sentiment clusters. Financial sources are identified, tiered, and verified against minimum thresholds. Each frame must meet minimum signal thresholds before a direction can be computed. Frames that don't meet thresholds are flagged for Insufficient Signal status. The relationship between financial signal strength and field sentiment distribution determines the directional call. Aligned signals produce Concordant; diverging signals produce Money Ahead or Mouth Ahead depending on which type leads. The CosentriQ team reviews all computed frames before a version is published to verify that each directional call is defensible and the signal is sound. No frame reaches the live map without passing this review. ## Coverage: Subpatterns and Industries The map currently covers five AI agent subpatterns across major industries. **Subpatterns:** | Subpattern | What It Covers | | ------------------------- | -------------------------------------------------------------------------------------------- | | Customer-Facing Workflow | AI agents in customer-facing contexts: support, sales assist, onboarding, service automation | | Developer Agents | AI in developer tooling: copilots, code review, test generation, CI/CD pipeline automation | | Horizontal Platforms | General-purpose AI platforms designed to operate across industries and functions | | Internal Ops | AI for internal business operations: HR, finance, legal ops, knowledge management | | Multi-Agent Orchestration | Systems that coordinate multiple AI agents for complex, multi-step tasks | **Industries currently in coverage:** Financial services, healthcare, legal, software and SaaS, retail and CPG, manufacturing, customer support and BPO, media and content, and government and education. Coverage expands with each map version as signal thresholds are met in new industry and subpattern combinations. Frames with insufficient signal are excluded from the published map. If your target category and industry combination doesn't appear as a frame, it means the CosentriQ team didn't have enough defensible signal to publish a directional call — not that the market doesn't exist. Check back after the next version refresh, or [contact the CosentriQ team](https://cosentriq.com) if you need intelligence on a specific combination.