Omni Analytics

1.1 Identity, product, and founder-market fit

Omni was founded in 2022 by a team reported to include Colin Zima, Christopher Merrick, and Jamie Davidson, all with Looker backgrounds. Current sources identify Zima as CEO and place the company in San Francisco. The product proposition is a governed semantic model that stores business definitions, relationships, and permissions, then applies them across dashboards, workbooks, spreadsheets, SQL, and AI queries. Omni says this model also governs external-agent access through its MCP server. That continuity between founder experience and product architecture is credible founder-market fit, but it is not proof of execution durability. The assigned evidence does not disclose a complete executive roster, board composition, succession plan, or independent assessment of AI accuracy and security.[CO001, CO002, CO003, CO004, CO014]

1.2 Capitalization, stage, and scale signals

Omni announced a $120 million Series C on April 23, 2026 at a $1.5 billion valuation. The transaction was led by ICONIQ, with Theory Ventures, First Round Capital, Redpoint Ventures, and GV participating, and included a $30 million employee tender offer. The announcement described a step-up from a $650 million valuation in March 2025 and claimed 4x year-over-year revenue growth. A current company-authored job listing says total capital raised is $217 million, but funding raised is not cash available and the public evidence provides no financing ledger, debt schedule, burn, or runway. Ramp reports a 6% adoption rate among organizations in its Business Intelligence vendor sample, up two percentage points year over year; that denominator is Ramp's measured vendor population, not Omni's customer count.[CO005, CO006, CO007, CO008, CO009, CO010]

1.3 Milestones, counter-signals, and diligence priorities

The evidence traces a founder transition from Looker, Omni's 2022 founding, a reported 2023 launch from stealth, a March 2025 valuation marker, the reported October 2025 Explo acquisition, and the April 2026 Series C. Current hiring and adoption observations extend the timeline without implying audited scale. Counter-signals come mostly from competitors or comparison platforms: Astrato highlights Omni's limited G2 review base as an enterprise procurement concern, Knowi says pricing is unpublished and sales-led, and RFP.wiki flags opaque pricing while ranking Omni seventh in its agentic-analytics comparison. Those are evaluation signals, not independent benchmarks of product quality. The assigned pool also does not establish whether material legal, regulatory, outage, or security incidents exist; absence from this bounded evidence set is not evidence of absence.[CO011, CO012, CO013, CO015, CO016]

1.4 Exhibits

2.1 Market boundary and buyer problem

The relevant market is governed, warehouse-connected analytics that lets teams define metrics once and reuse them across internal BI, self-service exploration, embedded customer analytics, and AI-assisted workflows. This boundary excludes raw data infrastructure, general-purpose LLMs, and visualization-only tools that do not own metric definitions or permissions. Omni's own documentation describes the product as combining a shared data model with SQL flexibility, while its AI material says the semantic layer defines metrics, dimensions, relationships, and business logic for consistent answers. The buyer problem is therefore not simply making charts; it is reducing metric drift and governance overhead while widening access to analysis. The principal substitutes are incumbent internal BI suites, specialist embedded analytics, a dbt-centered semantic layer paired with another BI tool, and the status quo of analyst-built SQL and spreadsheets.[CM001, CM002, CM004]

2.2 Demand signals and constrained sizing

Public market evidence is strongest on adoption intent and budget direction, not on a company-specific dollar opportunity. Futurum's 1H 2026 survey of 818 decision-makers at organizations with at least $100 million of annual revenue reports that 44.5% plan to increase semantic-layer spending within 24 months and 14.4% plan to newly adopt, or nearly 59% directing incremental budget. The same source reports 29.3% citing integration complexity and 24.6% citing lack of transactional write-back as infrastructure barriers. Gartner separately frames universal semantic layers as critical infrastructure by 2030, while Deloitte reports that 66% of surveyed organizations already report productivity or efficiency gains from AI but only 34% are truly reimagining the business. These figures establish a widening budget pool and adoption rationale, but they are not additive TAM layers and do not reveal Omni's obtainable share. The sizing table therefore preserves multiple lenses and explicitly marks the missing denominator.[CM005, CM006, CM007, CM008, CM009, CM010]

2.3 Buyer segmentation, adoption path, and constraints

The most plausible initial buyer is a data or analytics leader in a warehouse-first organization that needs governed metrics but cannot make every answer dependent on a central analyst team. Data teams own semantic modeling, permissions, and evaluation; business subject-matter experts and executives are the daily users; and product or engineering teams become budget owners when analytics is embedded into a customer application. Omni's partner description shows this workflow in practice: SMEs, analytic engineers, and executives use different parts of the same model, and one client migrated Tableau dashboards using Snowflake and Redshift connections. Ramp's October 2026 directional data places Omni at 6% adoption among organizations with a vendor in its Business Intelligence category, with 5% among enterprise companies; this is a vendor-panel adoption signal, not market share. Adoption is constrained by model-building discipline, integration complexity, procurement proof, custom pricing, and the risk that an incumbent suite or a focused semantic-layer product is already embedded in the workflow.[CM011, CM012, CM013, CM014, CM015]

2.4 Valuation relevance and unresolved sizing

For valuation, the attractive part of the market thesis is the possibility that one governed model becomes infrastructure for several budgets: BI, embedded analytics, and AI governance. The risk is that this breadth expands the competitive set and lengthens proof requirements. Independent category evidence points to skills, integration, governance, and measurable ROI as the gating variables, while competitor and partner material identifies custom pricing, limited public deployment signal, and incumbent alternatives as procurement friction. Omni-specific ARR, customer count, win rate, retention, contract value, and serviceable geography are not established by this pool. A defensible next step is to obtain cohort-level pipeline and renewal data, price by use case, implementation time, and the share of qualified opportunities that involve embedded or AI workloads before converting category percentages into a dollar forecast.[CM016, CM017]

2.5 Exhibits

The exhibits separate market definition, external demand signals, and the operating roles that must participate in an Omni purchase. Values are retained with their population, period, and limitations; unavailable figures remain null rather than being inferred. The sizing exhibit is intentionally a set of non-additive lenses: survey intent, forecast language, and directional vendor-panel adoption answer different questions. That distinction prevents a market narrative from turning broad category momentum into an unsupported Omni revenue forecast. The buyer map likewise distinguishes user, budget owner, and adoption trigger so that a diligence discussion can test who signs, who operates, and who realizes value.[CM005, CM011, CM016]

2.6 Exhibits

3.1 Competitive landscape and buyer alternatives

Omni's relevant competitive set is broader than direct semantic-layer vendors. Looker and Tableau represent incumbent enterprise platforms with semantic, embedded, and AI roadmaps; Power BI represents a broad-distribution suite; ThoughtSpot represents search- and AI-led analytics; Metabase represents a lower-cost open-source-to-commercial path; and Astrato is an adjacent modern analytics alternative. The status quo is a collection of warehouse SQL, spreadsheets, analyst-built dashboards, and internal tooling. Omni's own materials position one semantic model behind dashboards, embedded analytics, SQL, spreadsheets, conversational analytics, and MCP/API surfaces, while independent comparison material emphasizes that embedded buyers care about metric consistency, tenant boundaries, and AI governance rather than dashboard presence alone. This makes Omni's strongest wedge a governance-and-surface combination, not an unqualified feature lead.[CP001, CP002, CP003, CP004, CP005, CP006, CP007, CP008]

3.2 Capability, distribution, and commercial comparison

The capability comparison favors Omni where a buyer wants a common governed model spanning internal BI, embedded product analytics, and AI workflows. Looker and Tableau have materially stronger incumbent distribution and ecosystem leverage, while Power BI benefits from broad enterprise adoption and data connectivity. ThoughtSpot is a credible AI/search alternative, and Metabase can win on transparent entry pricing and open-source familiarity. Public list pricing is not comparable across the retained sources: Metabase's G2 listing shows Open Source at $0 and Starter starting at $85 per month for five users, while the Omni, Looker, Tableau, Power BI, and ThoughtSpot materials do not provide directly comparable realized prices. The absence of comparable Omni pricing is itself a diligence gap, not evidence that Omni is cheaper or more expensive.[CP009, CP010, CP011, CP012, CP013, CP015, CP016]

3.3 Moat durability and displacement risk

Omni's potential moat is the compounding value of a semantic model reused across many interfaces, plus governance controls and an AI agent layer that can operate on those definitions. Its security page claims ISO/IEC 27001:2022 certification, annual SOC 2 Type II audits, HIPAA compliance auditing, encryption, isolation, and least-privilege access controls. Its hiring evidence also shows active investment in agent harnesses, context management, retrieval, evaluation, and post-training. Those are meaningful capability signals, but they remain largely company- or job-posting-issued evidence. Against that, Looker, Tableau, and ThoughtSpot are adding governed conversational or agentic experiences, and broad suites can bundle analytics into existing relationships. Ramp reports 6% usage among organizations with a BI vendor in October 2026, with 5% enterprise adoption; this is a directional adoption signal, not market share or proof of switching. The durable question is whether Omni can convert cross-surface governance into lower deployment friction and repeatable distribution before incumbents commoditize the same workflow.[CP009, CP010, CP013, CP015, CP016]

3.4 Competitive diligence gaps

The public evidence does not establish realized pricing, win rates, churn or renewal performance by competitor, implementation effort, migration costs, feature-level performance benchmarks, or the share of Omni deployments using multiple surfaces. Reviews and vendor pages describe capabilities, but they do not provide a controlled head-to-head benchmark or customer-verified switching-cost study. Buyers should test semantic-model migration effort, row-level security and tenant isolation under production load, AI grounding and evaluation error rates, connector and embedded latency, administrator workload, and the commercial cost of replacing incumbent dashboards. Omni's roadmap velocity is promising, but the same pace can increase execution and platform-complexity risk if customers must continuously retrain workflows.[CP016]

3.5 Exhibits

4.1 Broad monetization surfaces, opaque contract economics

Omni publicly presents one governed semantic model across dashboards, workbooks, spreadsheets, SQL, embedded analytics, and AI workflows. That breadth creates several plausible commercial surfaces: the core analytics platform, customer-facing embedded delivery, and AI or agent usage. It does not establish that each surface is separately priced or reported as a revenue stream. Omni also says customers can create custom pricing tiers for AI products, while competitor-authored pricing reviews describe the platform as sales-led rather than governed by a stable public rate card. A customer review on AWS Marketplace separately flags difficulty navigating credit consumption. These observations support a recurring enterprise-software model with expansion potential, but not an estimate of realized ACV, revenue mix, discounting, or AI overage economics. The distinction matters because list architecture and product breadth are not evidence of recognized revenue or gross profit.[CI001, CI002, CI004, CI005, CI012]

4.2 Growth is visible; unit economics are not

The strongest current traction evidence comes from Fortune's independent reporting: ARR nearly quadrupled over the preceding year, Omni reached profitability for the first time in March 2026, and BambooHR was serving more than 100,000 users through Omni. Ramp's October 2026 vendor data separately reports 6% adoption among organizations using a business-intelligence vendor, up two percentage points year over year. These signals indicate commercial momentum and deployment scale, but they do not reveal absolute ARR, GAAP revenue, recurring versus services mix, NRR, churn, customer concentration, or sales efficiency. One month of accounting profitability is not evidence of sustained profitability or positive free cash flow. Direct-query warehouse execution and reliance on external LLM providers also expose cost drivers whose customer pass-through and margin effects remain unknown.[CI003, CI006, CI007, CI008, CI009, CI013, CI014, CI015]

4.3 A large round reduces financing pressure but does not establish runway

Omni announced a $120 million Series C at a $1.5 billion valuation in April 2026, with multiple retained sources reporting the same headline terms. The transaction also included a $30 million employee tender offer, so the headline funding amount should not be treated as cash available for operations without the financing documents and cash-flow statement. Public evidence does not disclose primary versus secondary proceeds, cash on hand, monthly burn, debt or credit facilities, project obligations, or the next-round trigger. The resulting verdict is positive on demand momentum and access to capital, but cautious on revenue quality, margin durability, and capital adequacy. Underwriting requires a data-room bridge from bookings to recognized revenue, cohorts and concentration, a gross-margin walk, and a fully diluted cash-and-obligations schedule before the financing can be translated into runway.[CI010, CI011, CI014, CI015, CI016]

4.4 Exhibits

5.1 Product and customer workflow

Omni is delivered as a direct-query business-intelligence and embedded-analytics platform. Its workflow begins with governed metrics and relationships, then lets business users ask natural-language questions, build dashboards, refine workbooks, or use SQL and spreadsheet-style formulas. The official product surface also lists Slack delivery, scheduled and conditional routines, APIs, and an MCP server. This creates a coherent path from a governed definition to an end-user answer, while preserving several interaction modes for technical and non-technical users. A partner implementation describes the same operating model as a layered system: raw database schema, governed data model, and ad-hoc workbook model. The practical buyer value is reduced metric disagreement and faster self-service; the diligence question is how much modeling and administration are required before those benefits appear in a real warehouse.[CE001, CE003]

5.2 Architecture and AI control plane

The strongest technical differentiator is the semantic layer as an execution boundary rather than a documentation layer. Omni says its AI generates semantic queries, which are translated to SQL against the warehouse, and that the model carries metrics, dimensions, joins, access controls, and AI-specific instructions. Its AI page says the coordinator plans actions, selects tools, executes queries, evaluates results, and can perform multi-step analysis. It also says Omni uses AWS Bedrock-hosted Claude models for most tasks and OpenAI models for advanced AI visualizations. These are credible product architecture claims, but they remain vendor descriptions: the public material does not provide independent benchmark results, latency distributions, evaluation datasets, or failure-rate evidence. A public MCP specification supports the interoperability premise, but not the security of any particular Omni deployment.[CE005, CE006, CE007, CE008, CE009]

5.3 Maturity, integrations, and dependencies

Omni exposes a broad integration and deployment surface: the official business intelligence page references APIs, an MCP server, Slack and email delivery, and bi-directional dbt integration. The official AI material describes AI Hub, reusable prompt-set evals, branch comparisons, and row-level filters. A public GitHub repository for an Omni Cursor plugin documents REST API access, API-key setup, and operations against the semantic layer, while also marking that plugin deprecated in favor of a successor repository. Ramp's proprietary panel reports 6% adoption among organizations using a business-intelligence vendor as of October 2026, but that category-level estimate does not establish production depth or workload scale. Together these signals show a real developer surface and market presence, but they also expose dependencies on cloud warehouses, model providers, APIs, and evolving agent tooling. No public source establishes uptime, support response, migration duration, or a complete release and deprecation policy.[CE010, CE012, CE013]

5.4 Trust, reliability, and implementation risks

Omni's control story is directionally strong: the company says row-level security and user attributes are enforced through the semantic layer, only necessary data is sent to external LLMs, and evals can compare prompt-set runs before production changes ship. However, the available evidence does not independently verify SOC 2, HIPAA, GDPR, incident history, or the effectiveness of those controls. A third-party product profile says setup and integration are required and that AI effectiveness depends on semantic-model quality; a partner account describes the layered operating model while noting that its partnership was new. An independent practitioner adds that semantic-layer projects surface unresolved ownership of metric definitions rather than solving that organizational problem automatically. The main product risk is therefore not a missing feature but implementation burden and operational dependence: customers must model business definitions, govern access, monitor AI behavior, and manage warehouse and model-provider changes. Buyers should require a live proof of concept with representative joins, permission tests, adversarial prompts, rollback evidence, and measured query cost before treating the platform as production-ready.[CE004, CE013, CE014, CE015, CE016, CE017]

5.5 Exhibits

6.1 Buyer, user, and payer segments

Omni addresses several related customer motions rather than one homogeneous buyer. Internal deployments sell governed analytics to data leaders and serve analysts, business teams, and executives through SQL, spreadsheets, point-and-click workflows, dashboards, and AI. Embedded deployments sell to SaaS product or data leaders, while their end customers become the users of reports and natural-language analytics. A third motion extends the same governed model into Slack, external agents, APIs, or MCP-enabled applications. This breadth may increase expansion potential because one semantic model can support multiple interfaces, but it also obscures the underlying revenue mix. Public sources do not split customers, bookings, or contract value by internal BI, embedded analytics, or AI use case, and they do not reveal whether pricing scales by creator, viewer, tenant, query, or another consumption unit.[CU001, CU002, CU003]

6.2 Adoption trajectory and named production proof

The strongest customer evidence is deployment-specific. Omni and independent coverage report that BambooHR launched Elite Analytics to more than 30,000 people in four months and later served more than 100,000 people, a materially stronger signal than a logo alone. Company-issued material also describes a three-month Cribl migration and a Guitar Center consolidation, while Omni's product guide says ActiveProspect rebuilt customer-facing dashboards in under two weeks. Partner and practitioner accounts add implementation texture: phData describes a Tableau migration for a client, and a consultant says Omni won an embedded ROI evaluation against Tableau and Looker. Ramp provides the broadest external adoption metric, estimating 6% adoption among organizations with a business-intelligence vendor in October 2026, up two percentage points year over year. That estimate lacks a visible sample denominator and is not interchangeable with customer count, ARR, or active usage.[CU004, CU005, CU006, CU007, CU008, CU009, CU010, CU011]

6.3 Durability, expansion, and concentration risk

The customer thesis is stronger on initial deployment than on durability. Omni's own surfaces describe an expansion loop from a governed model into self-service, embedded analytics, scheduled routines, and external AI workflows; named case summaries show that this path can reach customer-facing products. Yet the assigned evidence discloses no NRR, GRR, churn, renewal cohorts, contract length, account-level revenue, or top-customer share. Practitioner evidence also identifies friction that could limit expansion: a growing chart library, a fast-changing interface, and maintenance burdens from customer-specific customizations. AWS-hosted G2 reviews provide a positive 4.8/5 aggregate across 71 external ratings, but one October review says AI handled simple tasks while failing at dashboard construction. A competitor comparison raises additional procurement concerns, but it is competitor-issued rather than an independent benchmark. Investment diligence should therefore test cohort retention, paid expansion, production conversion, and concentration directly.[CU012, CU013, CU014, CU015, CU016]

6.4 Exhibits

7.1 Ranked risk profile

The highest residual risk is execution complexity. Omni is expanding from governed BI into agentic AI, MCP integrations, scheduled routines, and embedded workflows while a third-party comparison describes an Explo integration as mid-integration through 2026 and notes only 64 G2 reviews. That is not proof of failure, but it creates a measurable deployment and product-maturity risk for enterprise buyers. A second high risk is platform dependency: the company describes AWS Bedrock-hosted Claude models and public hiring material names Snowflake, BigQuery, Redshift, AWS, Databricks, and Bedrock in the operating stack. A change in model access, warehouse economics, or platform terms could affect reliability, gross margin, or roadmap control. A third risk is AI assurance. Omni says its AI Hub can run evals and that permissions and row-level filters are enforced through the semantic layer, but those are product claims rather than independent operating evidence. Investment monitoring should therefore prioritize release quality, migration completion, eval pass rates, material security events, dependency concentration, and customer renewal evidence.[CR001, CR002, CR003, CR004]

7.2 Regulatory and legal exposure

Omni's AI features can create regulatory exposure even if the platform is generally low risk in ordinary business-intelligence use. The European Commission says the AI Act's transparency rules apply from August 2026 and that the AI Office and member-state authorities become responsible for implementation and enforcement from 2 August 2026. The same source describes high-risk obligations including risk assessment, logging, documentation, human oversight, robustness, cybersecurity, and accuracy, with certain sensitive-use deadlines extending into 2027. Omni's public page says its AI uses external LLMs, semantic queries, row-level security, and result-set sharing for summarization; those mechanics make data-processing, disclosure, access-control, and customer-use-case scoping important diligence topics. NIST's framework is voluntary, so it is a useful control reference rather than a certification. The Cloud Security Alliance listing provides a CAIQ questionnaire context, but the retained text does not establish that Omni completed a certification or that controls were independently audited. No retained source establishes litigation, enforcement, or a material privacy incident, so those remain open questions rather than clean bills of health.[CR005, CR006, CR007, CR008]

7.3 Operational, security, and dependency risk

Omni has credible architectural mitigations but a broad operational blast radius. Its AI page says the semantic layer scopes topics and fields, applies user-attribute permissions and row-level filters, and lets customers inspect the SQL used for an AI response. It also says only necessary data is shared with external LLMs, distinguishing metadata from query results. These controls reduce the risk of unconstrained answers, but they do not show how often controls fail, how quickly incidents are contained, or whether customers receive complete audit trails. NIST's AI RMF emphasizes managing risks to individuals, organizations, and society and provides a voluntary framework for design, development, use, and evaluation; applying it could improve governance, but public adoption evidence is absent. The company also depends on several warehouses and external model infrastructure. A practical kill criterion is a material unauthorized data disclosure, repeated permission bypass, or sustained service degradation without a documented corrective-action trend. Procurement should request architecture diagrams, subprocessors, recovery objectives, penetration-test summaries, incident metrics, and dependency exit plans.[CR009, CR010, CR011, CR012]

7.4 Financial, partner, and thesis-break triggers

The financial-model risk is disclosure-driven rather than a known accounting loss. The retained public job material reports a $1.5 billion Series C valuation and $217 million raised, while the independent financing report describes a $120 million round, roughly 200 employees, nearly fourfold ARR growth, and first-time profitability in the prior month. These figures have different issuers and publication contexts, so they should not be treated as a complete cash, burn, or runway statement. A third-party adoption page reports 6% usage among organizations with a business-intelligence vendor, but it does not provide the denominator or customer economics needed to infer revenue quality. The main partner and concentration risks are dependence on cloud/model providers and the possibility that enterprise buyers require more deployment history, security evidence, or pricing transparency than public materials provide. The thesis should be downgraded if renewal or gross-margin evidence weakens, if the embedded integration misses its migration milestones, if external-model or warehouse costs compress economics, or if regulatory obligations materially delay deployments. Until those indicators are disclosed, the appropriate stance is to track with targeted diligence rather than assign a precise loss probability.[CR013, CR014, CR015, CR016]

7.5 Exhibits

8.1 Investment thesis and anti-thesis

The thesis is that Omni is building a coherent governed analytics platform rather than a standalone chatbot. The company describes AI operating on a semantic layer with governed metrics, joins, permissions, and reusable evaluation workflows, while third-party material describes the platform as combining self-service analytics, embedded delivery, and AI. A reported Series C at $1.5B and the hiring record's reference to a $120M round indicate substantial investor support, while a low-reputation third-party profile reports 4x annual revenue growth. These are financing and directional operating signals rather than proof of operating value. The anti-thesis is equally concrete: public revenue, ARR, retention, gross margin, burn, and customer concentration are not established by the assigned evidence. Competitive sources also identify custom pricing, a small review base, and integration or maturity questions. Without a denominator, a valuation multiple cannot be calculated responsibly, so the investment case depends on diligence rather than a forced buy call.[CV001, CV002, CV003, CV004]

8.2 Financing context and entry discipline

The strongest current price signal is a reported $120M Series C at a $1.5B valuation, alongside a separate company job listing that repeats the same valuation and round size. Those sources are not independent corroboration of a fair price: the article supplies a reported financing account, while the job listing is company-issued. The pool also contains a pricing benchmark that describes annual contract values from $15,000 for small teams to $150,000 or more for larger deployments, but it is anonymized benchmark context and not Omni revenue. An October 2026 public-software dataset places BI and analytics software at 3.2x EV/NTM revenue. That benchmark cannot be applied directly to a private financing valuation without a comparable NTM revenue basis, net cash or debt, growth, margins, and a public/private adjustment. Entry discipline should therefore require a verified latest cap table, primary round documents, preference stack, option-pool treatment, secondary/tender mechanics, current ARR or revenue, and a bridge from bookings to recognized revenue.[CV006, CV008, CV009, CV010]

8.3 Bull, base, and bear scenarios

The scenarios are decision ranges, not forecasts. The bull case requires verified sustained growth, strong retention, defensible gross margins, and evidence that semantic governance and embedded AI convert into durable expansion rather than feature parity. The base case assumes product traction continues but public financial disclosure remains thin, competition limits pricing power, and the reported financing mark is difficult to underwrite against a missing denominator. The bear case is triggered by a down round, weak renewals, commercialization friction, margin pressure from model or warehouse dependencies, or customer-facing analytics failures. The available evidence supports the direction of these sensitivities: Omni markets semantic-model grounding and evals, while independent and competitor material highlights custom pricing, limited review depth, ecosystem gaps, and the risk that semantic layers become a feature. Until the missing inputs are obtained, probability signals should remain qualitative and no scenario should be presented as a precise equity value.[CV011, CV012, CV013, CV014, CV015]

8.4 Recommendation, exit readiness, and diligence asks

Recommendation: track, with medium confidence and a stretched valuation stance at the reported $1.5B mark. This is not a judgment that the product is weak; it is a judgment that the price cannot be tested against public operating performance. A buy decision would require a verified revenue denominator, cohort retention, gross margin, net burn and runway, customer concentration, renewal evidence, and the exact Series C capitalization and preference terms. Exit readiness is also conditional: a credible path would require repeatable enterprise expansion, durable embedded-analytics adoption, clear AI quality and security controls, and enough audited or diligence-grade financial history for a strategic or public-market buyer to underwrite the business. Thesis-break triggers are a down round, material churn, a security or reliability event that weakens trust in governed analytics, inability to integrate embedded capabilities, or a valuation reset without corresponding operating growth.[CV002, CV003, CV016, CV017, CV018]

8.5 Exhibits