Celonis

1.1 Identity, headquarters, and current product positioning

Celonis describes itself in 2026 as the global leader in Process Intelligence and frames its homepage around the claim that "more than 1,400 companies around the world use Celonis to analyze, design, and operate AI-driven processes." The about page formalises the corporate identity with a dual-headquarter setup in Munich, Germany and New York City, USA, and reports a footprint of more than 20 offices worldwide alongside 3,000-plus employees and a self-reported aggregate of $6.5 billion of realised customer value. The product positioning is built around the Celonis Platform and its newly launched Celonis Context Model, which Celonis calls a dynamic, system-agnostic, real-time representation of a customer's entire business; this is the lens through which the company now sells Enterprise AI rather than process mining alone. The reusable identity for the rest of the report is therefore a Munich/New York-based, privately held process intelligence and enterprise-AI platform vendor, with explicit dependence on its own Context Model narrative and on a large installed base of SAP-anchored enterprise customers.[CO001, CO002, CO003, CO004, CO005, CO006]

1.2 Founders, leadership bench, and governance disclosure

The founder record is unusually clean for a private German software company. Wikipedia, Forbes, and an archived TU Munich School of Management profile all confirm that Celonis was founded in 2011 by Alex Rinke, Bastian Nominacher, and Martin Klenk as a spin-off from the Technical University of Munich, with the three meeting through a student consulting project for Bayerischer Rundfunk. Fifteen years later Rinke and Nominacher still run the company as co-CEOs and Klenk remains the named CTO, which means founder-market fit is intact and key-person dependence on the founding trio is structurally high. The executive bench around them has only recently thickened: Carsten Thoma, a Hybris co-founder and former SAP CX global president, was appointed President in August 2023, and Benoit Fouilland, formerly CFO of Criteo, Contentsquare, Firmenich, and Business Objects, became CFO on 1 December 2024 after a publicly unfilled CFO seat. Public board and committee disclosure remains thin, so governance depth beyond the founders, the President, and the CFO should be treated as a real diligence gap rather than silently papered over.[CO002, CO009, CO010, CO011, CO012, CO013]

1.3 Capital formation, investor map, and the unresolved 2024 round

Celonis's capital formation is well documented through 2022 and progressively murkier afterwards. TechCrunch and Reuters reporting confirms a $27.5 million Series A led by Accel and 83North in June 2016, followed by a $50 million Series B at a $1 billion valuation in June 2018 that made Celonis Germany's first software unicorn, and a $290 million Series C at $2.5 billion in November 2019. Forbes documents the June 2021 $1 billion Series D at $11 billion (NZZ called Celonis the country's first decacorn), and Bloomberg confirms the August 2022 Qatar Investment Authority-led Series D extension that fixed the most recent formal valuation at $13 billion. After that the public record thins: PitchBook lists a January 2024 Later Stage VC round with no disclosed amount, investors, or valuation, and the three main public databases now report meaningfully different lifetime totals - PitchBook $1.37 billion, Forbes "more than $1.6 billion," and Wikipedia approximately $1.77 billion. That spread and the unmarked 2024 round leave both ownership percentages and the question of any 2024-2026 down-round exposure unresolved without access to the cap table.[CO017, CO018, CO019, CO020, CO021, CO022]

1.4 Scale signals, ecosystem reach, and analyst validation

The strongest scale signals beyond valuation come from Celonis's customer roster, partner ecosystem, and analyst coverage rather than disclosed revenue, which the company does not publish. PitchBook's February 2025 snapshot records 3,353 total employees, and the about page now claims "3,000+" - both consistent with mid-thousands headcount and with a Fortune Future 50 No. 3 debut in 2025. Customer proof is dense: BMW Group has used Celonis for eight-plus years and has built a Process Mining Center of Excellence on the Process Intelligence Graph; the homepage and case-study library actively promote Fujitsu, Uniper, Standard Bank, PepsiCo, GE Healthcare, Deutsche Post DHL, Deutsche Telekom, Lufthansa Cargo, Carrefour, Lockheed Martin, Kraft Heinz, Uber, and Novartis. The partner side is equally dense: ServiceNow combines a strategic partnership with a direct equity investment, while IBM (watsonx and Consulting), Microsoft, AWS, Accenture, EY, and KPMG anchor the AgentC and Context Model rollout. Independent validation comes from Everest Group naming Celonis a Leader for the fifth consecutive year in 2024 and from Gartner's May 2026 Magic Quadrant Leader designation in the new Process Intelligence category.[CO024, CO028, CO029, CO030, CO031, CO032]

1.5 Milestone chronology and the SAP-adversary marker

The milestone record is rich enough that later chapters should not have to rediscover it. Celonis was founded in 2011, signed a reseller agreement with SAP in 2015, took its first institutional capital in 2016, crossed unicorn status in 2018, reached decacorn status with the June 2021 $11 billion Series D, and locked in its $13 billion mark with the August 2022 QIA-led extension. The acquisition cadence has been steady: Make/Integromat (October 2020, "over $100 million"), Lenses.io (October 2021, undisclosed), Process Analytics Factory (March 2022), Symbio/Symbioworld (November 2023), and most recently the 12 May 2026 agreement to acquire AI decision intelligence vendor Ikigai Labs alongside the launch of the Celonis Context Model. The single adverse marker that clearly belongs in the chronology is the breakdown of the SAP relationship: after SAP acquired direct competitor Signavio in early 2021, Celonis filed a US antitrust lawsuit against SAP in March 2025 alleging access restrictions to SAP system data, turning a long-time channel partner into an active legal adversary. That marker should carry into later risk work as a material structural risk to Celonis's SAP-anchored installed base, even though the primary docket URL was not captured in this run.[CO038, CO039, CO040, CO041, CO042, CO043]

1.6 Exhibits

2.1 Market boundary, adjacencies, and status-quo substitutes

Celonis should be bounded outward from its monetization surfaces rather than from every dollar labeled Enterprise AI. The company sells a Process Intelligence Platform whose three layers — Data Core, Celonis Context Model, and a Build experience for analyze/design/operate workflows — are anchored on extracting event logs from ERP, CRM, supply chain, and ITSM systems, then exposing the resulting process graph to humans and AI agents. That means the most relevant direct spend pools are process mining and process intelligence software, the adjacent process-analytics and task-mining markets, AI-decision-intelligence tooling pulled in via the announced Ikigai Labs acquisition, and the operational orchestration spend now flowing through agent platforms like Microsoft Copilot Studio, IBM watsonx Orchestrate, Amazon Bedrock Agents, and CrewAI. The boundary should exclude pure RPA seat licenses, generic data-warehouse spend, and foundation-model inference that never touches a Celonis process graph. Status-quo substitutes that Celonis must displace include spreadsheet-based lean/Six-Sigma audits, custom BI dashboards over ERP extracts, embedded SAP Signavio process modeling, UiPath and Microsoft process-mining add-ons sold inside broader automation suites, and open-source Apromore deployments inside academic or cost-sensitive enterprises. Gartner formally defines process mining as a technique for analyzing operational processes based on event logs, and IBM and UiPath now sell competing products under that same definition. The practical implication is that Celonis is no longer the only category leader label; multiple Leader designations exist across Gartner Magic Quadrant and Everest Group PEAK Matrix in 2024–2026.[CM001, CM002, CM003, CM004, CM005, CM006]

2.2 Sizing lenses, publisher contradictions, and demand backdrop

There is no single defensible TAM for Celonis. The closest analogue is the "process analytics" category that MarketsAndMarkets sized at USD 185.3 million in 2018 and projected to USD 1,421.7 million by 2023 at a 50.3% CAGR — a forecast that is now stale and well below current process-mining vendor revenue concentrations. More recent publishers (Grand View Research, IMARC, Fortune Business Insights, Precedence Research) publish process-mining software market reports, but most live behind paywalls or category-redirect pages on the public web, so absolute headline numbers are not independently verifiable in this run and are preserved here as evidence gaps. The outer envelope is IDC's 2025 FutureScape projection of USD 337 billion in worldwide AI-supporting technology spend rising past USD 749 billion by 2028, which Celonis itself cites as the tailwind for Enterprise AI orchestration but which clearly overstates Celonis-specific revenue pool. The narrowest lens is Celonis's own self-reported USD 6.5 billion in cumulative customer value to date — an output proxy, not a market size. IDC's own analyst Maureen Fleming and Forrester guests on Celonis-hosted content confirm process intelligence is becoming "foundational" for scaling agentic AI, while a 2023 Celonis-cited IDC outlook reframed process mining as "business value engineering." Each estimate uses a different boundary; collapsing them into one figure would mislead underwriting.[CM012, CM013, CM014, CM015, CM016, CM017]

2.3 Buyer segments, budget owners, and adoption path

The Celonis buyer map is broader than a single IT-platform budget. Verified customer stories and partnership content show the platform landing first with transformation and process-excellence offices (BMW established a Process Mining Center of Excellence in central IT), then expanding into supply chain (Uniper, Lufthansa Cargo, Fujitsu), finance and shared services (GE Healthcare unlocked USD 1.3 billion in free cash flow, Deutsche Post DHL HR operations), procurement (Carrefour, Deutsche Telekom award-winning procurement transformation), and IT/ITSM through the ServiceNow partnership (Siemens GBS named as first mover). Budget owners therefore include the CIO, CFO, CPO, COO, and increasingly the Chief AI Officer or transformation lead funding agent rollouts via Microsoft Copilot Studio, IBM watsonx Orchestrate, or Amazon Bedrock Agents. The adoption path Celonis itself describes runs from ERP/CRM data extraction to process discovery, then to KPI monitoring and execution apps, then to AgentC-based automation, then to the new Celonis Context Model as the operational substrate for Enterprise AI. Consulting partners (Accenture, IBM Consulting, KPMG, EY) gate large enterprise rollouts; Celonis's own "1,400+ companies" homepage claim, "1,500 customers" Forbes figure, and "5,000+ deployments" about-page claim show the buyer count is contested and may overstate net new logo penetration in the global mid-market.[CM023, CM024, CM025, CM026, CM027, CM028]

2.4 Growth drivers, adoption constraints, and valuation relevance

The demand backdrop is strong enough to justify continued category expansion. IDC says 2025 marks a shift from AI experimentation to reinvention; Celonis-cited Microsoft and Forrester commentary frames process intelligence as the missing operational context for AI agents; and Celonis's own "Four 2025 supply-chain predictions" piece and its 2026 tariff-management positioning argue that macro disruption (tariffs, supply-chain shocks, agentic-AI rollouts) increases willingness to fund process visibility. Celonis is also being designed into partner ecosystems as the data substrate for ServiceNow workflows, IBM watsonx, Microsoft Copilot, and AWS Bedrock — multiplying its surface area without proportionally growing its own sales force. Constraints are real and symmetric. Object-centric process mining (OCPM) is still novel; Celonis must educate buyers, as its own blog post on Everest Group concerns about OCPM acknowledges. The SAP relationship is now adversarial — Celonis sued SAP in March 2025 on antitrust grounds over data access — which directly threatens the SAP-event-log pipeline that historically powered the platform. Gartner's 2026 Magic Quadrant moved IBM, UiPath, SAP/Signavio, Microsoft, and Software AG/ARIS into Leader or near-Leader positions, raising the risk of price compression. Celonis also faces commoditization pressure from open-source Apromore (now part of Salesforce) and from automation-suite bundling. Layered on top, broad IT-spend governance, FinOps scrutiny, and uneven enterprise data quality slow conversion from PoC to multiyear ARR. For valuation, the consequence is that Celonis benefits from a credible secular tailwind, but durable upside depends on (a) winning data access fights, (b) retaining a defensible OCPM and Context Model lead, and (c) attaching to AI-agent budgets at materially higher net retention than the underlying process-analytics category alone would justify.[CM033, CM034, CM035, CM036, CM037, CM038]

2.5 Exhibits

3.1 Competitive landscape — direct, incumbent, adjacent, substitute, and status-quo paths

Celonis competes simultaneously in three overlapping arenas. The narrowest is dedicated process mining and process intelligence software, where the buyer is choosing between Celonis, SAP Signavio, IBM Process Mining, Microsoft Power Automate process mining, UiPath Process Mining, ABBYY Timeline, QPR ProcessAnalyzer, Software AG's ARIS, and the open-source Apromore stack. Gartner's published market definition frames this category as tooling that ingests event logs to discover, monitor, and improve operational processes. The second arena is bundled enterprise platforms: hyperscalers and incumbent application vendors fold process mining into broader Power Platform, watsonx, ServiceNow, or ERP suites and compete on bundle economics rather than mining depth. The third arena is status-quo substitutes — manual BPM, BI dashboards, custom event-log scripts, RPA-only point tools, and consulting-led six-sigma projects — which still absorb most enterprise process-improvement budget. The buyer can also build internally on cloud data platforms once event data is centralized, especially in firms that already pay for Databricks, Snowflake, or BigQuery. Substitution risk is therefore not just other process-mining vendors; it is the do-nothing path and the existing analytics stack.[CP001, CP002, CP003, CP004, CP005, CP006]

3.2 Direct and incumbent profiles — SAP Signavio, Microsoft, IBM, UiPath, and the next tier

SAP Signavio is the strongest direct competitor because SAP acquired Signavio in March 2021 for roughly 950 million euros and now ships it as a native part of SAP's business transformation suite, with privileged access to SAP ERP data Celonis publicly says it is being denied. Microsoft has folded process mining into Power Automate and Power Platform with explicit Copilot integration, which gives Microsoft a bundled GTM into every Power Platform tenant. IBM Process Mining is positioned alongside watsonx Orchestrate and IBM Consulting, and is sometimes also sold jointly with Celonis through the IBM Consulting partnership, so IBM is simultaneously a partner and a competitor. UiPath Process Mining and Task Mining ride RPA budgets and are tuned for "discover then automate" handoffs. The dedicated specialist tier is led by ABBYY Timeline (process mining plus task mining), QPR ProcessAnalyzer (Finnish vendor, MCP-native, partnered with Snowflake), and Software AG's ARIS family (BPM first, process mining second). Open-source Apromore is the most plausible commoditization vector: it offers enterprise process mining with academic roots and is positioned as a vendor-neutral alternative for governance and sovereignty-sensitive buyers.[CP010, CP011, CP012, CP013, CP014, CP015]

3.3 Capability, pricing, and packaging comparison

Public capability comparison is partial because most competitors publish marketing surfaces rather than full capability matrices, but enough is on the public web to compare anchor dimensions. Celonis pitches the Process Intelligence Graph, object-centric process mining (OCPM), the Celonis Context Model, and AgentC as a layered stack that connects raw event data to AI agents, with Gartner Leader placement reinforcing the breadth claim. SAP Signavio matches on ERP depth and now layers process intelligence inside SAP Business Suite, but is weaker on object-centric mining and multi-source coverage outside SAP. Microsoft and UiPath lean on automation handoffs; ABBYY and QPR lean on task mining plus AI root-cause analysis; ARIS leans on BPM/governance heritage. None of the major competitors publish standardized per-user or per-event pricing on their public sites, and Celonis itself does not publish list pricing, so total cost of ownership is opaque across the board. That pricing opacity is itself a competitive vector: bundled competitors (SAP, Microsoft, IBM) can present process mining as a zero-incremental-cost feature of an existing contract, while Celonis must justify a standalone enterprise subscription. Customers can multi-home process mining (some firms run Celonis alongside Signavio for SAP and Microsoft for Power Platform workloads), but that increases total switching cost only marginally because event data extraction is the dominant integration effort.[CP022, CP023, CP024, CP025, CP026, CP027]

3.4 Moat durability, distribution power, and adverse competitive evidence

The durable moats are the Process Intelligence Graph and Celonis Context Model as a semantic abstraction layer, AgentC for agent enablement, the AppStore/Marketplace of 1,000+ accelerators and apps, the customer reference base, and analyst placement (Gartner Leader in Process Intelligence Platforms 2026; Everest Group Leader and Star Performer for the fifth consecutive year). The most material adverse signal is the antitrust lawsuit Celonis filed against SAP in March 2025, which alleges SAP is leveraging its ERP dominance to lock Celonis out of native SAP data access — a complaint that, by definition, concedes SAP can constrain Celonis distribution regardless of product quality. Hyperscaler bundling is a second material vector: Microsoft pulling process mining into Power Automate, and IBM pulling it into watsonx Orchestrate, both lower the price umbrella under which Celonis must justify a separate platform contract. Open-source Apromore and the IEEE Task Force on Process Mining (which standardized the XES event-log format) reduce vendor lock-in at the data-format layer. The offsetting structural lock-in is integration depth and customer-specific data models: once a customer has built hundreds of object-centric process models and AgentC agents inside Celonis, ripping them out has real switching cost even if the underlying mining algorithms commoditize.[CP031, CP032, CP033, CP034, CP035, CP036]

3.5 Exhibits

4.1 Revenue model, monetization surfaces, and what is publicly observable about traction

Celonis monetizes a single integrated platform — the Celonis Platform built around the Data Core, the new Celonis Context Model, and Build experiences — sold to large enterprises as a process-intelligence and Enterprise AI subscription. The company itself states that "more than 1,400 companies" use Celonis, while the Forbes company profile cites 1,500 customers and the company's about page references 5,000+ enterprise deployments as a separate metric, so even the simplest top-of-funnel number is reported three different ways depending on the publisher. No list pricing is published on the company's own surfaces; in contrast to public-cloud comparators such as UiPath, Pegasystems and Appian (all SEC registrants), Celonis treats pricing as a sales-led, contract-by-contract conversation. That means every revenue, ARR, gross margin, take-rate, average contract value, and discount-band figure for Celonis is private, and any chapter that pretends otherwise is over-fitting. What the public record does support is a multi-stream revenue surface: core process-mining and process-intelligence subscriptions, add-on Enterprise AI capabilities including the new Celonis Context Model and AgentC, ecosystem integrations with SAP, ServiceNow, Microsoft, IBM watsonx and Databricks that are revenue-relevant via co-sell, and a small services and partner-led implementation layer. Customer-disclosed value (e.g., GE Healthcare's $1.3B free cash-flow lift) is marketing proof, not Celonis revenue.[CI001, CI002, CI003, CI004, CI005, CI006]

4.2 GTM motion, customer cohorts, and the unit-economics proxies that actually exist

Celonis does not publish CAC, payback, sales-cycle length, ACV, NRR, or quota productivity. The right diligence stance is therefore to triangulate from public proxies: the scale and quality of the named customer book, partner-distribution coverage, and headcount as a sales-capacity input. PitchBook's February 2025 snapshot pegs Celonis at 3,353 total employees, while the company itself says "3,000+" employees across 20+ offices, implying a meaningful field organisation. The named customer book — BMW, GE Healthcare, Deutsche Post DHL, Deutsche Telekom, Lufthansa Cargo, Carrefour, Lockheed Martin, PepsiCo, Kraft Heinz, Uber, Uniper, Fujitsu, Novartis, Standard Bank, Siemens GBS, AstraZeneca, ASOS, Dell and Johnson & Johnson — is institutional-grade and skews to multi-year, multi-process accounts where land-expand economics are highly likely but not measurable from outside. Distribution leverage from the ServiceNow strategic partnership (in which ServiceNow took a financial investment) and from SAP, IBM, KPMG, EY, Accenture, AWS and Microsoft partner motions is the most plausible explanation for Celonis sustaining enterprise scale without disclosed revenue. Independent profiling services Sacra, Tracxn and CB Insights either gate Celonis financial disclosures behind paid tiers or show no useful numerics on their free pages, which is itself an evidence signal: even specialist private-market trackers do not have an underwriteable view of Celonis revenue today.[CI009, CI010, CI011, CI012, CI013, CI014]

4.3 Cost structure, margin drivers, and what public comparables imply for Celonis

With no audited financials, the only honest cost-structure read is via mechanism and comparable filings. Celonis's cost base is dominated by R&D and field sales for a Munich/New York dual-HQ organisation of roughly 3,000–3,353 people, plus hyperscaler infrastructure consumed when Celonis runs the Data Core and Context Model on AWS, Databricks and Microsoft Fabric. The December 2024 appointment of Benoit Fouilland — formerly CFO at Criteo (which he took public on Nasdaq), Contentsquare, Firmenich and SAP AG — was explicitly framed by the company as a move to "drive the next phase of growth," widely read as IPO-preparation discipline. Public software comparables suggest a realistic envelope: Pegasystems trades as a profitable mid-cap enterprise software company whose financial summary is publicly disclosed via the SEC, Appian files 10-Ks that show how a low-code enterprise platform reports gross margin and operating losses, and UiPath has documented consumption-vs-subscription pricing pressure since its 2021 IPO. None of those filings substitute for Celonis data, but they bound the question: a process-intelligence platform with Celonis's customer concentration should be capable of 70–80%+ software gross margin if delivery is genuinely software, and materially lower if the partner-services tail is large. The Ikigai Labs acquisition (announced 2026-05-12) adds R&D headcount and MIT-licensed patents — a capability buy, not a revenue-accretive deal — which suggests Celonis is still investing ahead of the curve rather than harvesting margin.[CI019, CI020, CI021, CI022, CI023, CI024]

4.4 Capital stack, dilution lens, and forward capital adequacy

The Company Overview chapter walks the round-by-round chronology; this section instead reframes the same facts through a capital-stack and adequacy lens. Publicly reported total raised diverges meaningfully across authoritative sources: PitchBook's February 2025 profile lists $1.37B raised, Forbes's company profile says "more than $1.6 billion," and Wikipedia's March 2026 article cites "approximately $1.77 billion." The most likely explanation is that PitchBook excludes the ServiceNow strategic investment, certain secondaries, and an undisclosed January 2024 "Later Stage VC" deal whose amount is not public. The last formal valuation mark is the August 2022 $13B Series-D extension led by Qatar Investment Authority, which followed the June 2021 $11B Series D that made Celonis Germany's first decacorn — meaning Celonis has not raised at a fresh valuation in nearly four years, a long mark-life by late-stage software standards. There is no public cash balance, no monthly burn figure, and no runway disclosure, so capital adequacy must be inferred. Positive signals include the CFO hire framed around growth discipline, ongoing acquisitions (Make/Integromat, Lenses, PAF, Symbio, Sailfin, and the pending Ikigai Labs deal) which imply available capital, and the No. 3 ranking on the 2025 Fortune Future 50. Negative signals include no public up-round since 2022, persistent IPO rumours that have repeatedly slipped, the broader 2023 enterprise-software layoff wave that hit ~140 Celonis roles, and the SAP antitrust lawsuit filed by Celonis in March 2025 which converts the company's largest historical distribution channel into an active legal counterparty.[CI029, CI030, CI031, CI032, CI033, CI034]

4.5 Financial verdict, evidence gaps, and underwriting diligence path

The verdict is that Celonis presents as a well-capitalised, enterprise-anchored, late-stage private software company with a credible product/market story and an identifiable path to IPO, but with a public financial profile that is materially thinner than peers of comparable scale. The chapter cannot responsibly assert ARR, revenue growth, gross margin, operating margin, cash, burn, or runway because no verified primary source discloses them; the company-website "+$13B valuation" and "$6.5B realised customer value" are marketing aggregates rather than financial statements. The most important diligence blockers are therefore: (1) the exact 2025 and trailing-twelve-month revenue and ARR, with bookings vs. recognised revenue split; (2) audited gross and operating margin; (3) a current cash, runway and burn schedule including any venture-debt covenants; (4) clarity on the January 2024 PitchBook "Later Stage VC" event, including any flat or down mark; (5) any mutual-fund mark-to-market disclosures (e.g., from Fidelity or T. Rowe Price holders) implying private-market re-rating since the 2022 $13B reference; and (6) IPO timing and use-of-proceeds plan, including whether Celonis intends to clear a US listing in 2026 as widely rumoured. Adverse risks worth pricing in are the SAP antitrust litigation outcome (which could either accelerate Celonis distribution or strand it), persistent down-round risk relative to the $13B mark, and a cumulative absence of disclosure that resembles other late-stage private companies which subsequently revised valuations sharply at IPO.[CI029, CI030, CI039, CI041, CI042, CI043]

4.6 Exhibits

5.1 Product scope: the Celonis Platform as a single integrated process-intelligence and Enterprise AI surface

Celonis sells a single platform — the Celonis Platform — built around the Data Core, the new Celonis Context Model launched on 12 May 2026, Build experiences (Studio, Apps, Action Flows, Marketplace) and the AgentC agent layer first introduced at Celosphere on 23 October 2024. The customer-visible job is object-centric process mining (OCPM): connect to enterprise systems (SAP, Oracle, ServiceNow, Salesforce, Workday, custom databases), build a digital twin of operations across objects and events, find inefficiencies and value leakage, and then orchestrate change either through Action Flows automations or, increasingly, through AI agents that consume the same Context Model. The company's platform page frames the surface as 'industrialize Enterprise AI' rather than as a stand-alone process-mining tool, and the public docs site (docs.celonis.com) confirms that the same Data Core, PQL query language and PI Graph metadata layer underpin both the human-facing Studio and the agent-facing AgentC and CCM surfaces. Marketplace.celonis.com hosts partner-built apps, connectors and industry packages, and the company maintains a public GitHub organisation (github.com/celonis) plus a PyCelonis SDK referenced in docs, so the platform is genuinely developer-addressable even though no self-serve pricing is published.[CE001, CE002, CE003, CE004, CE005, CE006]

5.2 Architecture: Data Core, the Context Model, OCPM, PQL, PI Graph and run-on-anywhere connectivity

Architecturally Celonis layers the platform as ingestion + Data Core + Context Model + Build/Apps + AgentC. The Data Core is the storage and compute layer that ingests structured event and object data from source systems via 100+ connectors, normalises it into an object-centric event log, and exposes it through PQL (Process Query Language) and the PI Graph metadata surface. On top of the Data Core, the Celonis Context Model — described in the 12 May 2026 press release as the platform's 'digital twin of operations' — adds a semantic layer that makes process knowledge consumable by external AI agents and copilots. CCM run-on-anywhere connectivity is explicitly enumerated in the press release: AWS, Databricks, Microsoft Fabric, Snowflake (announced as 'coming soon'), and Oracle. Build experiences (Studio, Apps, Action Flows) sit above Data Core/CCM and give business analysts a low-code/no-code surface to build process applications, automations, and metrics; the same surface is documented at docs.celonis.com. OCPM is itself documented in Celonis's own 'What is OCPM?' blog and is the academic frame underlying patent EP4468169A1 on object-centric data model graphs filed by Celonis SE inventors Nolle, Glampedakis, Doumanis, Matthaei, Utz, and Newman. Process-tree discovery is in turn covered by patent EP4481646A1 on an approximate inductive miner.[CE003, CE009, CE010, CE011, CE012, CE013]

5.3 AgentC and the Enterprise AI integration surface: October 2024 vs May 2026 lineup

AgentC is Celonis's agent-layer SDK and integration surface, first announced at Celosphere on 23 October 2024. The October 2024 AgentC press release enumerates a much narrower ecosystem than the May 2026 Context Model launch does, and the chapter preserves that delta as a conflicting evidence signal rather than resolving it. The October 2024 release explicitly names Microsoft Copilot Studio, IBM watsonx Orchestrate, Amazon Bedrock and CrewAI as the launch ecosystem; the May 2026 CCM press release additionally names Anthropic Claude Cowork, Databricks Agent Bricks, Microsoft Agent365 / Copilot, Oracle OCI Enterprise AI, and (via the Ikigai Labs acquisition) MIT-licensed decision intelligence patents. The agreed acquisition of Ikigai Labs on 12 May 2026, with Devavrat Shah named Chief Scientist Enterprise AI and MIT becoming a Celonis shareholder, is the most concrete capability addition since AgentC launched. From a deployment posture this means a Celonis customer in 2026 can in principle plug the Context Model into a Bedrock agent, a watsonx Orchestrate flow, a Copilot/Agent365 surface, or a CrewAI multi-agent workflow — but evidence for production depth (vs. listed integration) is uneven across partners, which is recorded as an evidence gap. AgentC also requires AI agents to consume PQL through the Data Core/Context Model, which is the lock-in mechanism that differentiates Celonis from horizontal RPA or LLM tooling.[CE005, CE017, CE018, CE019, CE020, CE021]

5.4 Ecosystem: SIs, Marketplace, GitHub, customer-proof for deployment depth

Celonis is highly SI- and partner-led on the deployment side. IBM Consulting publicly states it has more than 2,000 certified Celonis consultants and frames the IBM–Celonis partnership as a co-sell and co-deliver motion across SAP-heavy enterprises; that scale of practitioner certification is itself a developer-signal proxy in a market where official GitHub stars and npm downloads understate real adoption. ServiceNow, SAP (despite the active Celonis antitrust filing), Microsoft, AWS, Databricks, Accenture, KPMG and EY round out the SI/co-sell layer documented across the Celonis press, partner pages, and the Celonis Next 2025 partner summit. Marketplace.celonis.com (JS-only page, captured here as accessStatus js-only) curates partner-built apps, connectors and industry packs that extend the Data Core into vertical workflows. Customer-proof for production depth is documented through named case-study blogs (Kraft Heinz, Lockheed Martin, Uber, GE Healthcare, PepsiCo), and the Celonis 'Where enterprise tech goes next' and 'Object-centric process mining' blogs argue the thesis for why OCPM and CCM together are the differentiating architecture rather than the agent integrations alone. The Celonis GitHub organisation (github.com/celonis) is live but hosts a small number of public repos relative to typical developer-tool vendors, and the PyCelonis Python SDK is documented at docs.celonis.com rather than published on PyPI's public index page (pypi.org/project/pycelonis returned 404 in this run) — recorded as an evidence gap.[CE006, CE007, CE008, CE025, CE026, CE027]

5.5 Trust, IP and roadmap: granted patents, undiscoverable trust page, and the 2026 release pipeline

Celonis's strongest public technical-IP signals are two granted European patents indexed on Google Patents: EP4481646A1, on an approximate inductive miner for process-tree discovery (assignee Celonis SE, filed 19 June 2023, published 25 December 2024, inventors van Detten, Leemans, Schumacher); and EP4468169A1, on an object-centric data model graph (priority 23 May 2023, published 27 November 2024, inventors Nolle, Glampedakis, Doumanis, Matthaei, Utz and Newman). The patents directly back the OCPM and process-discovery architecture rather than being marketing IP. On trust and compliance, however, the obvious URLs celonis.com/trust/ and celonis.com/security/ returned 404 in this run (preserved as evidenceGaps and an accessStatus row in the trust/quality table); the Celonis platform page asserts enterprise security and compliance but does not enumerate specific certification scopes (SOC 2 Type II, ISO 27001, ISO 27018, HIPAA, FedRAMP, GDPR, PCI) at a machine-checkable surface here. The 2026 roadmap is anchored to the CCM launch, the Ikigai Labs close, expansion of AgentC ecosystem partners, and the Solution Suite line (Supply Chain, Finance, Front Office, Sustainability) referenced across celonis.com and the Celonis Next 2025 partner summit; the Gartner Peer Insights page on Celonis (JS-only) is a reputation signal but not a roadmap source. Roadmap depth beyond 2026Q3 is not publicly enumerable and is logged as an evidence gap with diligence paths.[CE014, CE015, CE016, CE031, CE032, CE033]

5.6 Exhibits

6.1 Customer base composition, count contradictions, and visible buyer map

Celonis sells almost exclusively to large enterprises, and the public buyer map skews heavily to the Global 2000 across manufacturing, automotive, consumer goods, life sciences, financial services, telecom, logistics, energy, public sector, and pharma. The company's own About Us page says "more than 1,400 companies" use Celonis, while the Forbes Cloud 100 profile lists 1,500 customers and the company separately references "5,000+ deployments" — a different metric that almost certainly counts modules, use cases, or business-unit instances rather than logos. None of these numbers can be reconciled externally because Celonis has never published a full customer list, a region split, a vertical split, or a revenue-band breakdown. The /customers/ URL on celonis.com redirects to a curated /solutions/stories carousel that surfaces only a few dozen logos at a time. Geographically the disclosed reference set leans European DACH and US large-cap, with selective Asia-Pacific (Fujitsu) and African (Standard Bank) exposure, matching the Munich/New York co-headquartered GTM footprint and the SAP/Microsoft/IBM partner channels documented in chapter 5. Buyers cluster in CFO/finance, COO/operations, CIO/CTO, CPO/procurement, and supply-chain functions; the user surface is then process-engineering and analyst teams, with business-process owners as the indirect beneficiaries. The payer is almost always a central transformation or shared-services budget, which is why most public references describe enterprise-wide rollouts (BMW, Novartis, PepsiCo) rather than department-only pilots.[CU001, CU002, CU003, CU004, CU005, CU006]

6.2 Named customer proof — outcomes, freshness, and depth of deployment

The strongest public asset Celonis owns is a deep library of named customer references with quantified operating outcomes. BMW Group is the deepest reference: the March 2024 press release confirms a partnership that began in 2016, a dedicated BMW Process Mining Center of Excellence, and a Group-wide rollout aimed at making Celonis available to every employee — anchored by joint development on the Process Intelligence Graph. GE Healthcare publicly attributes a $1.3 billion free-cash-flow increase in a single year to its Celonis-led transformation. Deutsche Telekom reports 100% electronic POs, 92% no-touch invoice processing, €12 million in savings, and a >96% on-time payment rate after automating procurement with Celonis. Kraft Heinz cites a 30% reduction in overdue payments using process mining. Lufthansa Cargo runs Object-Centric Process Mining over ~2 million Air Waybill records and >80 million activities a year, originally to optimize revenue accounting. PepsiCo reports "millions unlocked" in accounts receivable, an 86% reduction in rejected sales orders, and >1,000 hours/year saved in accounts payable. Standard Bank rebuilt its cross-border payments process with Celonis (and partner ProcessLab), pushing straight-through processing above 90% and cutting turnaround from up to 55 hours to "a handful." Lockheed Martin tied Celonis into its $4B "1LMX" multi-year process-transformation program. Carrefour, Uber, Uniper (with Microsoft), Cosentino, Fujitsu, Siemens GBS, Saint-Gobain, and Novartis (with Accenture, since 2018, across 15+ use cases) round out a reference set whose depth and 2024–2026 freshness compare favorably with any private enterprise SaaS vendor.[CU010, CU011, CU012, CU013, CU014, CU015]

6.3 Retention, satisfaction signals, and what is still missing

Public retention proof for Celonis is qualitative, not quantitative. The company has never disclosed NRR, GRR, logo churn, contract length, ARR cohorts, or weighted average remaining duration; sell-side coverage in chapter 4 inherits the same opacity. What can be observed are review-site signals and multi-year customer continuity. TrustRadius rates Celonis 9.5 out of 10 across 252 reviews and shows best-in-class scores in process discovery (above 8.9 category average), process mining algorithms (above 8.8), and data connection (above 8.7). G2 shows 4.5 out of 5 across 302 reviews with a published time-to-implement of four months. FeaturedCustomers aggregates 194 Celonis customer references with 102 testimonials, 44 case studies, 48 customer videos, and a 4.7/5.0 rating based on 3,589 reference ratings. Multi-year continuity itself is evidence of renewal: Novartis has been live since 2018, BMW since 2016, and Lufthansa Cargo and Deutsche Telekom both feature in multi-vintage case studies. Recurring negative themes in the same review surfaces are also clear: a steep learning curve for PQL and advanced analyses, performance lag on very large datasets, integration complexity for non-standard systems, and — from a Novo Nordisk reviewer on TrustRadius — a request for a more mature Lifecycle Management model for regulated industries. None of this substitutes for audited GRR/NRR; an investor must request the underlying cohort and renewal data directly.[CU023, CU024, CU025, CU026, CU027, CU028]

6.4 Expansion logic, partner channels, and concentration risk

The land-and-expand pattern visible in Celonis case studies is consistent: a single procure-to-pay or order-to-cash pilot in finance or shared services, followed by horizontal expansion into supply chain, IT, HR, audit, and customer service, then anchored by a customer-built Center of Excellence (BMW, Novartis, Standard Bank all match this arc). Partner channels — SAP (since 2015), ServiceNow, Microsoft, IBM Consulting + watsonx, AWS, Accenture, KPMG, EY — extend distribution and procurement surface; many large references (BMW, Uniper, Carrefour, Novartis) were originally co-sold or co-implemented. Concentration risk is harder to read: with no top-10 ARR disclosure, no segment mix, and a heavy DACH/Fortune 500 reference skew, it is plausible that a small number of platform accounts represent an outsized share of revenue. The SAP relationship has become structurally ambiguous following Celonis's March 2025 US antitrust lawsuit against SAP (covered in chapter 7), which simultaneously threatens a long-standing customer-access channel and potentially exposes shared accounts to SAP-Signavio replacement pressure. The customer-side risk is therefore not a logo-quality problem but a channel-dependence and concentration-disclosure problem that only private diligence can close.[CU032, CU033, CU034, CU035, CU036, CU037]

6.5 Adverse signals — review complaints and customer-quoted gaps

Adverse customer signals are real but bounded. The G2 review pool surfaces consistent friction on four points: (i) advanced features such as PQL and custom analyses have a steep learning curve; (ii) performance can lag on very large or complex datasets; (iii) industry-specific templates are still thin outside the most common verticals; and (iv) integration with smaller or custom-built tools requires technical expertise. The TrustRadius Novo Nordisk Associate Digital Director review makes a more pointed regulated-industry critique: "the platform's capabilities are evolving faster than we can validate them to our regulated processes, which limits how we are allowed to use Celonis," and requests a differentiated environment / mature Lifecycle Management model. Brocacef (a 501–1,000-employee Dutch pharma logistics firm) flags slow guidance on object/catalog modeling and an over-reliance on generic formula examples on the website. Independently, no public mass-churn event, no enterprise lawsuit from a customer, no large G2/TrustRadius rating collapse, and no publicly named failed deployment surfaced in 2024–2026 — the most material public adverse event of the period is Celonis's own action against SAP (a partner/competitor) rather than any customer revolt. The aggregate picture is a complex, expensive platform that requires sophisticated buyers but does not appear to be losing them.[CU029, CU030, CU031, CU039, CU040, CU041]

6.6 Exhibits

7.1 Legal and regulatory risk is anchored by the SAP antitrust lawsuit and live EU AI Act exposure

Celonis's most material legal exposure as of May 2026 is the antitrust complaint it filed against SAP in the US District Court for the Northern District of California on 4 March 2025, alleging that SAP weaponised its ERP-data position — specifically through RISE with SAP and Grow with SAP migration packaging — to choke off third-party access to the customer data Celonis needs to operate. Bloomberg Law's antitrust desk, The Information, and Bloomberg Law's IP desk all reported that the case is now in active discovery, and a parallel Celonis patent suit against SAP escalates the legal posture from commercial dispute to multi-front litigation against its single largest historical distribution partner. CourtListener returns 98 hits on "celonis" across US federal dockets, indicating that the SAP matter is not the only active proceeding. Beyond US litigation, the EU AI Act entered force on 1 August 2024 with phased obligations: penalty and general-purpose-AI duties from 2 August 2025, the bulk of high-risk obligations from 2 August 2026, and full applicability by 2 August 2027. Because the October-2024 AgentC launch and the May-2026 Celonis Context Model are explicitly positioned as the substrate for enterprise AI agents acting on customer ERP data, Celonis is plausibly both a provider of AI systems and a deployer for customers in regulated verticals (banks, insurers, healthcare, automotive), which puts it inside the documentation, training-data-disclosure, and risk-management duties the Act imposes on GPAI and high-risk uses. German federal financial supervisor BaFin and the EDPB-aligned national DPAs will separately scrutinise data-access and AI-governance arrangements that Celonis ships into banks and insurers. These are not theoretical filings — they convert directly into contract friction, slower regulated-vertical GTM, and reserve-and-disclosure questions an IPO prospectus would need to answer.[CR001, CR002, CR003, CR004, CR005, CR006]

7.2 Operational, security, partner-concentration, and execution risk all transmit through SAP and the AI rollout

Celonis's operational risk surface is dominated by partner concentration, not by obvious uptime or breach events (no public security-incident register, breach notification, or status page was retrievable from primary Celonis sources during this run — itself a diligence gap). The most concentrated dependency is SAP: Celonis grew up as an SAP Startup Focus participant in 2012, signed a global reseller agreement in 2015 (sold as "Celonis Process Mining by SAP"), and built its installed base around customers running SAP ECC and S/4HANA. SAP's own ownership of Signavio (a 2021 acquisition explicitly positioned as a process-mining/process-modelling competitor) plus the RISE/Grow with SAP data-access dispute mean that the single counterparty that historically sourced and distributed Celonis is now simultaneously an active legal adversary, an aggressive competitor, and a gatekeeper to roughly half the installed base's underlying ERP data. Beyond SAP, competitive pressure stacks: Microsoft is bundling Power Automate Process Mining inside the Microsoft 365 / Power Platform estate; UiPath pushes Process Mining inside its existing automation footprint; IBM ships its own process-mining product and simultaneously partners with Celonis via watsonx; Appian and Pegasystems carry adjacent low-code workflow narratives. The AgentC / CCM rollouts compound execution risk: launching enterprise-AI substrate while simultaneously closing the Ikigai Labs acquisition (announced 12 May 2026), absorbing PAF (2022), Symbio (2023/2024), Integromat/Make (2020), Lenses.io (2021), and Banyas (2019) creates real integration, talent, and product-management load against a 3,000+ headcount that already absorbed two rounds of layoffs. Talent risk is live: Glassdoor was unreachable to verify employee sentiment in this run, Sifted and Handelsblatt coverage of the 2023 and 2024 reductions returned 404s through the standard fetch path, and the company has had no confirmed public up-round since 2022.[CR012, CR013, CR014, CR015, CR016, CR017]

7.3 Financial and valuation risk is shaped by a four-year mark gap, an IPO delay, and undisclosed concentration

Celonis has not printed a public up-round since the QIA-led Series D extension at $13B in August 2022 — a gap of nearly four years by the runDate of this report. Wikipedia and PitchBook record a January 2024 "Later Stage VC" round but with no disclosed valuation, which Crunchbase, NZZ, and PitchBook coverage collectively leave consistent with — but not proof of — down-round risk. Total raised is itself reported inconsistently across primary trackers ($1.37B PitchBook, $1.6B+ Forbes, ~$1.77B Wikipedia), which a serious IPO prospectus would need to reconcile. Reuters reported in September 2025 that Celonis was exploring a 2026 US IPO, and Bloomberg followed in June 2025 (cached as a 404 via the reader fallback) with reporting that management told staff the IPO had been put on hold amid market volatility — a sequence consistent with The Information and Handelsblatt coverage of the IPO preparation timeline. Layoffs reinforce the burn-versus-mark picture: Sifted and Handelsblatt reported a ~10% reduction in mid-2023 (~140 roles against a then-~3,000 headcount), BigTechWire flagged a follow-on restructuring in February 2024, and the company subsequently hired Benoit Fouilland as CFO (1 December 2024) and Carsten Thoma as President (August 2023) — moves consistent with IPO-grade financial-controls and operating-discipline preparation but not a substitute for audited disclosure. Customer-concentration data is entirely undisclosed against a 1,400–1,500 logo base whose SAP-running profile already overlaps the antitrust exposure described above; an unexpected top-customer churn into SAP Signavio or Microsoft would be visible only after the fact. The combination — premium private mark, no fresh validation round, paused IPO, layoffs, opaque concentration — is what makes valuation, rather than product, the largest live risk.[CR026, CR027, CR028, CR029, CR030, CR031]

7.4 Governance, geography, and people risk plus monitorable thesis-break triggers

Celonis remains co-CEO led by founders Alex Rinke and Bastian Nominacher, with co-founder Martin Klenk as CTO. Co-CEO structures concentrate decision rights inside a personal relationship that has not been publicly stress-tested through a recession, a public-market filing, or a hostile-acquirer event. The recently added President (Carsten Thoma, ex-Hybris/SAP) and CFO (Benoit Fouilland, ex-Criteo) provide professional balance but are themselves new-in-seat. Geographic concentration is real: Munich plus New York City headquarters, European DACH-leaning reference base, German legal domicile and BaFin/BSI regulatory environment, and a German IPO ecosystem that has been historically thin make Celonis structurally exposed to European macroeconomic, geopolitical (Ukraine, energy, sanctions), and capital-markets conditions in a way US-domiciled peers like UiPath, Databricks, or Snowflake are not. IP risk is two-sided: the Celonis-vs-SAP patent suit asserts Celonis IP, but it also implicitly puts Celonis IP into public-domain scrutiny by counter-claim. The most actionable output of this chapter is therefore a set of monitorable kill criteria: an adverse SAP-suit ruling, an explicit down-round, a second IPO postponement past 2027, a third round of layoffs, a Gartner downgrade from Leader, or a public customer defection from Celonis to SAP Signavio would each independently compress the bull case. The recommendation is to treat the SAP-suit docket, BaFin and Bundeskartellamt registers, the EU AI Act implementing-act tracker, layoffs.fyi, and CourtListener as standing diligence surfaces — none of which the company controls — and to require management to map AgentC / CCM into provider vs deployer categories under the AI Act before underwriting any premium to the 2022 mark.[CR023, CR024, CR025, CR033, CR036, CR037]

7.5 Exhibits

8.1 Investment thesis, anti-thesis, and recommendation

Celonis is the category-defining process-intelligence vendor, with a 2026 Gartner Magic Quadrant leader position, a 1,400+ enterprise customer base spanning BMW, Deutsche Post DHL, GE Healthcare, Lockheed Martin, PepsiCo, and Standard Bank, and a fresh AI narrative built on the Celonis Context Model launch and the announced Ikigai Labs acquisition on 2026-05-12. The thesis is that process mining is the durable substrate for enterprise AI agents, that Celonis has a multi-year head start, and that a 2026 US IPO would crystallize the value already implied by the August 2022 QIA-led $13B round. The anti-thesis is unforgiving: the last formal mark is almost four years old, Sacra estimates roughly $700M ARR for 2024, public process / workflow comps (Pegasystems, Appian, UiPath) have de-rated to 4–6x revenue, Bloomberg reported in June 2025 that Celonis told staff IPO plans were on hold, Sifted and Handelsblatt documented two waves of layoffs in 2023 and 2024, and SAP's antitrust counter-narrative caps the partner upside. On public evidence the recommendation is track, with valuation stance stretched and a buy call deferred until either an IPO discount or audited disclosure resets the denominator.[CV001, CV002, CV003, CV004, CV005, CV006]

8.2 Last formal mark, secondary signals, and disclosure quality

The anchor mark is the QIA-led Series D extension announced on 2022-08-23 at a reported $13B post-money, preceded by the June 2021 Series D at $11B reported by Forbes. Total raised is itself contested across primary trackers: PitchBook cites roughly $1.37B, Forbes lists $1.6B+, and Wikipedia aggregates a figure closer to $1.77B; the contradiction is material because dilution and preference stacks change every scenario downstream. Public secondary marks for Celonis are not transparently disclosed: Forge Global lists the company on its private-market platform but the public-facing insight page returns boilerplate rather than executed pricing. Mutual-fund marks have not been re-published since the 2023 layoff cycle, and The Information's December 2025 and Reuters' September 2025 reports describe IPO exploration without committing to a price band. The Bloomberg June 2025 staff note that the IPO was on hold for market volatility is the most adverse signal on this axis. Disclosure quality is therefore weak relative to the late-stage software peer set: there is no audited revenue, no margin disclosure, no cap-table waterfall, and no formal valuation refresh since the 2022 round.[CV012, CV013, CV014, CV015, CV016, CV017]

8.3 Comparable companies, multiple band, and scenario logic

The defensible comp universe spans three layers. The closest workflow / process peers — Pegasystems, Appian, and UiPath — trade at roughly 4–6x trailing revenue per Wikipedia market-cap snapshots cross-checked with their SEC 10-K listings, anchoring a conservative valuation. The high-growth data-infrastructure layer — Snowflake and Datadog — runs at roughly 14–17x revenue and sets a realistic mid-band ceiling, given Celonis' AI-agent framing and 1,400+ enterprise footprint. ServiceNow, with a roughly $200B market cap on about $11–12B revenue, acts as the aspirational long-run anchor showing what scaled, profitable enterprise workflow software can earn in the public market. Bessemer's BVP Nasdaq Emerging Cloud Index data and the State of the Cloud 2024 narrative both document that the median public cloud multiple compressed materially from 2021 highs, which is exactly the reset SaaStr and Meritech-derived commentary describe. Applied to a Sacra estimated ARR of about $700M for 2024, the bear case (5x) lands near $3.5B, the base case (12x using the data-infra blend) near $8.4B, and the bull case (18x with an AI-agent premium) near $12.6B — still below the 2022 $13B sticker.[CV023, CV024, CV025, CV026, CV027, CV028]

8.4 Recommendation, thesis-break triggers, and final diligence asks

The committee call is track with medium confidence, valuation stance stretched, risk rating high, and an explicit ask for IPO-window discipline before any incremental commitment. The thesis-break triggers are concrete: a down-round or secondary print below $9B, ARR growth slipping below 25%, an adverse outcome in the SAP antitrust dispute filed in March 2025, a third layoff wave, or a 2026 IPO being formally postponed for more than two consecutive quarters. The diligence asks that would move the call are equally concrete: audited revenue, ARR, gross margin, and free-cash-flow bridges; the cap-table preference stack including the QIA-led Series D terms; a quantified read on SAP-related ACV concentration and reseller revenue at risk; and a verified mutual-fund mark or executed Forge / EquityZen secondary print to triangulate the current implied valuation. Until those land, public evidence supports patience rather than purchase. The KPI dashboard and the valuation-sensitivity chart make the trade-offs explicit, and the recommendation callout below restates the stance so the report-meta summary can cite it directly without re-deriving it.[CV038, CV039, CV040, CV041, CV042, CV046]

8.5 Exhibits