The Arion Enterprise AI Atlas: Taxonomy

Companion to the Native & Embedded Classification Methodology | v0.13 | September 2026 | Arion Research

The Atlas is organized by two complementary lenses over the same set of products. Part I (Structure) sorts products by how they are built, Native or Embedded, and the software category they belong to: the supply-side view. Part II (Business Function and Use Case) sorts the same products by who has the problem and what the product does about it: the demand-side view. Every product carries a placement in both. One version and one version history cover the whole taxonomy; the version history is at the end.

Part I. Structure: Native and Embedded

1. Purpose

The methodology defines how a solution is classified (the five tests, the decision rule, the edge-case precedents). This document defines where it lands: the full set of category cells in both panels, each with a boundary statement, what belongs and what does not, and representative products already placed there. Published together, the two documents make every placement on the map transparent and contestable.

Categories are data rows, not schema. They evolve with the Atlas, and every add, split, or merge is recorded in the edition delta and in the version history at the end of this document.

2. How to read the taxonomy

The Atlas has two panels, and each panel is organized on its own principle:

The Native panel is organized as a stack, read bottom to top: the physical and model foundations at the base, the data and orchestration layers above them, a trust layer across the middle, and applications at the top. A product sits in the layer that describes what it primarily is.

The Embedded panel is organized by the host software category the AI capability lives inside. The question is not what the AI does but what established product it augments. A CRM copilot sits under CRM regardless of how sophisticated the AI is.

The unit is always the product or solution, not the vendor, so one vendor can appear in several cells and in both panels. When a placement is genuinely contested, the five tests decide it and the rationale is recorded on the product.

Agentic capability is an attribute, not a category. Every product carries an agentic level (0 None, 1 Assistive, 2 Agentic, 3 Autonomous); products at level 2 or above show an agentic tag on the map. The tag cuts across every category and is never itself a category. See the methodology, Section 6.

Native and Embedded pairings. Several capabilities appear in both panels by design, and the pairing is the point: the standalone AI-first product is Native, while the same capability delivered inside an incumbent host is Embedded. The clearest pairs are Enterprise Search & Knowledge (Native) with Search & Knowledge Platforms (Embedded); Data Analysis (Native) with Analytics & BI Platforms (Embedded); Software Development & Coding (Native) with Development & DevOps Platforms (Embedded); Finance & Operations (Native) with ERP & Finance Suites (Embedded); HR & Talent (Native) with HCM & Talent Suites (Embedded); AI Security (Native) with Security Platforms (Embedded); the Native Content & Creative and Sales & Marketing cells with the Embedded Marketing Technology category; the Native Customer Experience & Service cell with the Embedded CRM and Contact Center categories; the Native Manufacturing & Supply Chain cell with the Embedded Procurement and Supply Chain & Logistics categories; and, across the vertical layer, the Native Vertical Application cells with the Embedded Industry Clouds category. Reading a pair together shows the framework working: similar capability, different classification.

3. Native panel

Organized as a stack, bottom to top. Counts are products currently placed in the cell; "Roadmap" marks a defined cell not yet populated.

3.1 Infrastructure & Compute

AI Compute & GPU Cloud (8 products). Accelerated compute and specialized cloud whose reason to exist is training and inference workloads. Boundary: in scope are neoclouds and GPU clouds built for AI demand; out of scope are general-purpose hyperscaler regions, which appear through their embedded vendor entries rather than as neoclouds. Representative: CoreWeave, Nebius, Crusoe, Lambda GPU Cloud.

Inference & Model Serving (10 products). Platforms whose core job is serving, scaling, and optimizing model inference. Boundary: in scope are managed inference and model-serving clouds; out of scope is raw GPU capacity with no serving layer (that is Compute) and model labs that happen to offer an API (those are Models). Representative: Together AI Inference, Fireworks, Baseten, GroqCloud.

3.2 Models

Frontier Model Providers (8 products). Labs whose primary offering is state-of-the-art foundation models and their APIs. Boundary: in scope is the proprietary flagship model or API sold as the product; out of scope are the same vendor's open-weight releases (those go to Open & Specialized) and their application products (those go to the relevant app cell). Representative: Claude for Enterprise, ChatGPT Enterprise, Google Gemini API, Amazon Nova.

Open & Specialized Models (10 products). Open-weight, domain-specific, and small-model providers. Boundary: in scope are open-weight families and specialized or small models distributed as the product; out of scope are proprietary frontier flagships (Frontier) and fine-tuning services (Data Layer). Representative: Meta Llama, Mistral open models, IBM Granite, Microsoft Phi.

3.3 Data Layer

Vector & Retrieval Infrastructure (8 products). Databases and services purpose-built for embedding storage and retrieval. Boundary: in scope are dedicated vector and retrieval engines; out of scope are incumbent databases that added a vector feature, which are Embedded under Data Platforms & Clouds (the "vector great divide"). Representative: Pinecone, Weaviate, Qdrant, Zilliz Cloud (Milvus).

Data Pipelines & Integration for AI (5 products). Pipelines, connectors, parsing, and context preparation built specifically for model consumption. Boundary: in scope are tools whose purpose is turning source material into model-ready context; out of scope are general ETL platforms that added AI features (Embedded, Data Platforms & Clouds). Representative: Unstructured, LlamaCloud (LlamaParse), Reducto, Extend.

Labeling & Synthetic Data (12 products). Data annotation, curation, human feedback, and synthetic-data generation for AI. Boundary: in scope are products whose output is training, alignment, or evaluation data; out of scope are analytics products that consume data rather than produce it for models. Representative: Scale Data Engine, Snorkel Flow, Labelbox, Surge AI.

3.4 Orchestration & Agent Infrastructure

Agent Platforms & Frameworks (15 products). Platforms for building, deploying, and running AI agents. Boundary: in scope are horizontal agent build-and-run platforms; out of scope are agents sold as extensions of an incumbent suite (Embedded) and single-purpose vertical agents (the relevant app cell). Representative: LangGraph Platform, CrewAI AMP, OpenAI Agents SDK & AgentKit, Amazon Bedrock AgentCore, xAI Grok Bot, StackAI.

Model & Agent Gateways (Roadmap). Routing, caching, cost control, and policy layers sitting between applications and models or agents. Boundary: in scope are dedicated gateway and routing products; a candidate move here is OpenRouter, currently placed in Inference. Not yet populated.

Interoperability & Protocols (Roadmap). Products built around agent and tool interoperability standards such as MCP and A2A. Boundary: in scope are products whose value is the interoperability layer itself, not merely support for a protocol. Not yet populated.

Agent Memory & Context (Roadmap). Memory, state, and long-term context management infrastructure for agents. Boundary: in scope are dedicated memory and context stores for agents, distinct from vector databases used only for retrieval. Not yet populated.

3.5 Trust, Safety & Governance

Evals & Observability (9 products). Evaluation, testing, tracing, and monitoring of models and agents. Boundary: in scope are dedicated eval and observability products; out of scope are general APM tools that added an LLM view and governance tooling focused on policy rather than measurement. Representative: Arize AX, LangSmith, Langfuse, Braintrust.

AI Security (8 products). Security for and against AI: model protection, prompt-attack defense, and agent security. Boundary: in scope are independent AI-security pure plays; out of scope are AI-security modules absorbed into an incumbent security platform (Embedded, Security Platforms). Representative: HiddenLayer, Zenity, Noma Security, Straiker. (Lasso Security is held on watchlist.)

AI Governance, Risk & Compliance (7 products). Policy, audit, model inventory, and regulatory compliance for AI systems. Boundary: in scope are standalone AI-GRC platforms; out of scope are governance modules that ride an incumbent data or privacy platform (Embedded, Data Platforms & Clouds or Privacy & GRC). Representative: Credo AI, Holistic AI, Cranium, Modulos.

3.6 Horizontal Applications

Applications usable across industries. Each pairs with an Embedded counterpart where an incumbent ships the same capability.

Software Development & Coding (7 products). AI-native coding assistants, agents, and software-engineering platforms. Boundary: in scope are standalone AI-first developer tools; the Embedded counterpart is GitHub Copilot and similar, under Development & DevOps Platforms. Representative: Cursor, Claude Code, Devin, Replit Agent.

Enterprise Search & Knowledge (7 products). Model-mediated search and knowledge products across enterprise content. Boundary: in scope are AI-first search and knowledge assistants; the Embedded counterpart is pre-LLM search vendors under Search & Knowledge Platforms. Representative: Glean, Hebbia Matrix, Perplexity Enterprise Pro, Moveworks.

Productivity & Meetings (5 products). AI-native productivity, meeting, and communication products. Boundary: in scope are standalone AI-first tools; the Embedded counterpart is productivity-suite copilots under Collaboration & Productivity Suites. Representative: Granola, Fireflies.ai, Otter.ai, Superhuman.

Content & Creative (8 products). Generative content creation for enterprise marketing, design, and media. Boundary: in scope are AI-first content and media generation products where the models are the product; out of scope are creative features inside an incumbent marketing or creative suite (Embedded, Marketing Technology). Representative: Jasper, Synthesia, HeyGen, ElevenLabs, Adobe Firefly. (Adobe Firefly is the model family itself, so it is Native here; the embedded generative and agentic AI inside Adobe's Creative Cloud apps is a separate, Embedded entry.)

Customer Experience & Service (8 products). AI-native customer service and experience agents. Boundary: in scope are standalone AI-first service agents; the Embedded counterparts are CRM & Customer Experience Platforms and Contact Center & CCaaS Platforms. Representative: Sierra, Intercom Fin, Decagon, Ada.

Sales & Marketing (6 products). AI-native selling, outreach, and marketing-execution products. Boundary: in scope are AI-first GTM and sales-agent tools; out of scope are AI features inside an incumbent CRM or martech suite (Embedded, CRM & Customer Experience Platforms or Marketing Technology). Representative: Clay, Rox, 11x (Alice), Qualified (Piper AI SDR). (AI-native marketing-execution candidates below the enterprise bar, held on watchlist: Regal AI, Tofu, Inflection.io, SuperAGI, Ploy.)

Data Analysis (5 products). AI-native analytics and data-work products. Boundary: in scope are AI-first analysis tools where model-mediated analysis is the product; the Embedded counterpart is Analytics & BI Platforms, where incumbent BI suites that added AI belong. Representative: Hex, Julius AI, Zenlytic, Definite, Athenic. (Watchlist native candidates below the enterprise bar: Fabi.ai, Querio.)

HR & Talent (6 products). AI-native recruiting and talent products. Boundary: in scope are standalone AI-first HR and recruiting tools; the Embedded counterpart is HCM & Talent Suites. Representative: Paradox (Olivia), Juicebox (PeopleGPT), Eightfold, Metaview.

Finance & Operations (6 products). AI-native finance, accounting, and operations products. Boundary: in scope are AI-first finance and ops tools; out of scope are AI features inside an incumbent ERP or finance suite (Embedded, ERP & Finance Suites). Representative: Basis, Rillet, Campfire, Numeric.

3.7 Vertical Applications

Applications built for a single industry. The classification is still Native; the axis is the served industry. Each cell pairs with the Embedded Industry Clouds category, where incumbent industry suites that added AI are placed.

Legal (8 products). AI-native products for legal work. Boundary: in scope are AI-first legal products; out of scope are horizontal tools merely used by legal teams, and incumbent legal-research platforms that added AI (Embedded, Industry Clouds). Representative: Harvey, Legora, Robin AI, CoCounsel Legal.

Healthcare & Life Sciences (7 products). AI-native products for clinical, research, and life-sciences work. Boundary: in scope are AI-first clinical documentation, decision-support, imaging, and discovery products; out of scope are incumbent EHR/health-cloud suites that added AI (Embedded, Industry Clouds). Representative: Abridge, Ambience Healthcare, OpenEvidence, Aidoc.

Financial Services (9 products). AI-native products for banking, insurance, and investment work. Boundary: in scope are AI-first FS products (research, underwriting, compliance, claims); out of scope are FS incumbents that added AI. Representative: Rogo, Sixfold, Shift Technology, Zest AI, Fisent BizAI.

Manufacturing & Supply Chain (6 products). AI-native products for industrial and supply-chain work. Boundary: in scope are AI-first machine-health, inspection, negotiation, supply-chain-graph, and decision-intelligence products where the models are the product; out of scope are incumbent industrial or supply-chain platforms that added AI (Embedded, Supply Chain & Logistics or Industry Clouds). Representative: Augury, Pactum, Altana, Interos, Aera Decision Cloud.

Public Sector & Education (5 products). AI-native products for government, defense, and education. Boundary: in scope are AI-first products built for public-sector and education missions; out of scope are incumbent government/education suites that added AI (Embedded, Industry Clouds). Representative: Polimorphic, Govini Ark, Scale Donovan, MagicSchool AI, Anduril Lattice.

Other Verticals (6 products). AI-native vertical products in industries that do not yet warrant a dedicated cell. A holding category; a cell graduates out of it once it has enough members to stand alone. Current members span real estate, construction, retail, energy, and agriculture. Representative: EliseAI (real estate), Buildots (construction), Constructor (retail), Urbint (energy), Carbon Robotics (agriculture), Everseen Vision AI Platform (retail).

4. Embedded panel

Organized by host software category. The AI capability lives inside an established product that predates it and functions without it. Counts are products currently placed; "Roadmap" marks a defined cell not yet populated.

CRM & Customer Experience Platforms (12 products). AI capability embedded in established CRM and CX suites. Boundary: in scope are AI copilots and agents sold as part of a CRM or CX suite; out of scope are AI-first service agents with no host suite (Native, Customer Experience & Service). Representative: Agentforce, Microsoft Dynamics 365 Copilot, Oracle AI for Fusion CX, HubSpot Breeze, Sprinklr AI+, Salesforce Agentforce Commerce. (Salesforce Agentforce Commerce / Commerce Cloud is placed here as the nearest Salesforce host; a dedicated Commerce cell is a future split candidate. Gainsight Atlas held on watchlist. Pega, previously placed here, moved to Process Automation & BPM.)

ERP & Finance Suites (14 products). AI capability embedded in established ERP, financials, close, and enterprise-planning suites that predate the AI and run without it. Boundary: in scope are copilots and agents inside an incumbent ERP or finance suite; the Native counterpart is Finance & Operations, and AI-first finance startups belong there, not here. Representative: SAP Joule for Finance, Oracle AI Agents for Fusion Cloud ERP, Microsoft Dynamics 365 Finance & Supply Chain Copilot, NetSuite AI, Sage Copilot, plus finance-suite incumbents BlackLine (close), Anaplan (planning), and Workiva (reporting and disclosure). Note: Workday Illuminate spans finance but is mapped once under HCM & Talent Suites, not duplicated here.

HCM & Talent Suites (16 products). AI capability embedded in established HCM, HR, and payroll suites that predate the AI and run without it. Boundary: in scope are AI copilots, assistants, and agents inside an incumbent HCM/payroll suite; the Native counterpart is HR & Talent, where AI-first recruiting and talent startups belong, not here. Spans enterprise suites (Workday Illuminate, Oracle AI Agents for Fusion Cloud HCM, SAP SuccessFactors Joule, ADP Workforce Now / ADP Assist, UKG Pro / Bryte AI, Dayforce Co-Pilot, Infor HCM, Darwinbox) and modern mid-market platforms (Rippling, Paycom, Paylocity, HiBob, Deel, Personio, isolved, Paychex). Note: SMB-anchor HRIS (BambooHR, Gusto) and Paycor (consolidating under Paychex) are held on watchlist rather than mapped.

Collaboration & Productivity Suites (9 products). AI capability embedded in established productivity, collaboration, and document platforms. Boundary: in scope are suite copilots; the Native counterpart is Productivity & Meetings. Representative: Microsoft 365 Copilot, Atlassian Rovo, Notion AI, Zoom AI Companion, Box AI, Adobe Acrobat Studio, Egnyte AI Agents, Asana Agentic Work Management, Slack (Slack AI / Agentforce in Slack). (Box and Egnyte are content-cloud/ECM, Adobe Acrobat Studio is document/PDF productivity, and Asana is work management; all are placed here as the nearest host category. A dedicated Content Management cell is a future split candidate if the segment deepens.)

ITSM & IT Operations (2 products). AI capability embedded in established IT service and operations platforms. Boundary: in scope are AI agents and control planes inside an ITSM platform. Representative: ServiceNow AI Agents, ServiceNow AI Control Tower.

Contact Center & CCaaS Platforms (5 products). AI capability embedded in established contact-center-as-a-service (CCaaS) and customer-engagement platforms. Boundary: in scope are AI features inside a CCaaS platform; the split from CRM & CX keeps voice and contact-center incumbents distinct from CRM suites. Representative: NiCE CXone Mpower, Genesys Cloud AI, Five9 Genius AI, Talkdesk AI (Ascend), Twilio CustomerAI. (Twilio is CPaaS, adjacent to CCaaS; placed here as the nearest host category.)

Analytics & BI Platforms (22 products). AI capability embedded in established analytics, BI, and analytics-automation platforms that predate the AI. Boundary: in scope are AI copilots, assistants, and agents inside an incumbent BI/analytics suite; the Native counterpart is Data Analysis, where AI-first analytics tools (Hex, Julius) belong, not here. Spans hyperscaler/enterprise BI (Power BI Copilot, Gemini in Looker, Amazon Q in QuickSight, Oracle Analytics Cloud, SAP Analytics Cloud), independents (Tableau, Qlik, ThoughtSpot Spotter, Strategy/MicroStrategy, Domo, Sisense, Sigma), legacy suites (IBM Cognos, SAS Viya, Spotfire, Pyramid), analytics automation (Alteryx), product analytics (Amplitude), open-source BI (Metabase), plus GoodData, Zoho Analytics, and Adobe Customer Journey Analytics.

Marketing Technology (22 products). AI capability embedded in established marketing platforms: marketing automation, customer engagement, CDP/journey orchestration, and marketing clouds that predate the AI and run without it. Boundary: in scope are AI copilots, assistants, and agents inside an incumbent martech suite; the Native counterparts are Content & Creative (generative creation) and Sales & Marketing (AI-first GTM execution), where AI-native martech startups belong, not here. Spans marketing-cloud/suite incumbents (Salesforce Marketing Cloud Next / Agentforce for Marketing, the Adobe Experience Cloud cluster — Marketo Engage, Experience Platform Agent Orchestrator, Journey Optimizer, and GenStudio for Performance Marketing — Oracle Fusion Marketing, SAP Engagement Cloud / Emarsys, Microsoft Dynamics 365 Customer Insights, Zeta Global's Athena) and the customer-engagement and email/messaging field (Braze Sage AI, Klaviyo, Iterable, Attentive, Bloomreach Loomi, MoEngage Merlin, CleverTap Clever.AI, Insider Sirius AI, Optimove OptiGenie, ActiveCampaign, Intuit Mailchimp). Adobe Creative Cloud's embedded AI (the generative and agentic features inside Photoshop, Premiere, Illustrator, and Express) is also placed here as the nearest host category, since these creative-production tools are used primarily by enterprise marketing and creative teams; a dedicated Embedded Creative & Design Tools cell — pairing with the Native Content & Creative panel — is the natural future home if that segment deepens (it would also hold future Canva/Figma embedded AI). Webflow (AI Assistant / AI Site Builder) is also placed here as the nearest host: it is a visual web-development and web-experience platform whose core predates the AI, marketed to marketing teams; a dedicated Web Experience / DXP cell (which would also hold Sitecore, Optimizely, Contentful, and similar) is another future split candidate. Webflow is mapped at Level 1 today; its agentic platform, Source by Webflow, is in limited research preview and would raise it to Level 2 at GA. Note: HubSpot Marketing Hub's AI is Breeze, mapped once under CRM & Customer Experience Platforms and not duplicated here; Marigold's enterprise assets (Sailthru/Selligent) moved to Zeta Global in 2025 and are covered through the Zeta entry.

Search & Knowledge Platforms (3 products). AI capability embedded in established enterprise search and knowledge-management platforms that predate the model era. Boundary: in scope are pre-LLM search vendors that layered on generative answers; the Native counterpart is Enterprise Search & Knowledge. Representative: Coveo, Sinequa. (Guru held on curation.)

Data Platforms & Clouds (12 products). AI capability embedded in established data platforms, warehouses, lakehouses, and data-management clouds. Boundary: in scope are incumbent data platforms that added vector search, AI governance, agents, or model features; this is the Embedded side of the vector great divide. Representative: Databricks Mosaic AI, MongoDB Atlas Vector Search, pgvector, IBM watsonx.governance, Snowflake Cortex AI, Informatica IDMC (CLAIRE AI), Salesforce Data Cloud. (Informatica joined Salesforce in Nov 2025; its IDMC data-management cloud with CLAIRE AI and Salesforce Data Cloud's grounding/vector layer both sit here.)

Security Platforms (22 products). AI capability embedded in established security platforms: SIEM, security operations, endpoint/XDR, and threat-management suites that predate the AI and run without it. The deepest single cell in either panel. Boundary: in scope are AI copilots, assistants, and agentic SOC analysts inside an incumbent security platform, plus AI-security modules absorbed into one; the Native counterpart is AI Security, where independent AI-security pure plays (HiddenLayer, Zenity, Noma, Straiker) belong, not here. Spans the enterprise SIEM/SecOps field (Microsoft Sentinel with Security Copilot, Google SecOps with Gemini in Security, CrowdStrike Falcon Charlotte AI, Splunk Enterprise Security, Palo Alto Cortex XSIAM, IBM QRadar watsonx, Securonix EON, Exabeam Nova, Rapid7 Incident Command, Elastic Security, Datadog Cloud SIEM Bits AI, Sumo Logic Mo Copilot, Fortinet FortiSIEM, Check Point Infinity AI Copilot, OpenText ArcSight Aviator), the endpoint/XDR field (SentinelOne Purple AI, Trend Vision One / Trend Companion, Sophos AI Assistant, Bitdefender GravityZone / CLAIRE, Tanium AI), and the earlier network/AI-defense modules Cisco AI Defense and F5 AI Guardrails. Note: Palo Alto's Prisma AIRS and Check Point's Lakera sit in the Native AI Security cell (distinct products from these SecOps platforms); Microsoft Defender XDR uses the same Security Copilot already mapped via Sentinel and is not double-counted; mid-market/MSP security platforms (Graylog, Coralogix, Cynet, Panther, ManageEngine Log360, Trellix Wise, Cybereason, WithSecure, ESET, and others) are held on watchlist.

Privacy & GRC Platforms (2 products). AI-governance capability embedded in established privacy-management and enterprise GRC platforms. Boundary: in scope are privacy and GRC incumbents that added AI governance; the Native counterpart is AI Governance, Risk & Compliance. Representative: OneTrust AI Governance, Drata AI Agent Governance.

Procurement (8 products). AI capability embedded in established procurement / source-to-pay (S2P) and spend-management suites that predate the AI and run without it. Boundary: in scope are AI copilots and agents inside an incumbent procurement/S2P suite; the Native counterpart is Finance & Operations for AI-first spend tools. Representative: Coupa Navi, SAP Ariba (Joule Agents), GEP Quantum Intelligence, JAGGAER JAI, Ivalua IVA, Zip AI, LightSource AI Workforce, Workday Contract Lifecycle Management (CLM, powered by Evisort). (Workday CLM is a contract-lifecycle-management product placed here as the nearest host cell, since contract management sits in the S2P space; a dedicated Contract Management / CLM cell — which would also hold Icertis, Ironclad, SirionLabs, DocuSign CLM, and SAP/Coupa CLM — is a future split candidate if that segment deepens. Workday's separate Contract Intelligence product, also built on the Evisort acquisition, is a candidate to add alongside it.) (Zip is the most AI-forward and closest to the Native boundary; kept Embedded on Test 1 because the intake-to-procure orchestration core predates the agents. Renamed from "Procurement & Supply Chain" in v0.9 when Supply Chain & Logistics was split out.)

Supply Chain & Logistics (6 products). AI capability embedded in established supply-chain planning, execution, and visibility platforms that predate the AI and run without it. Boundary: in scope are AI copilots and agents inside an incumbent SCM planning, WMS/TMS/OMS, or real-time-visibility platform; the Native counterpart is Manufacturing & Supply Chain, where AI-first machine-health, negotiation, and supply-graph startups belong, not here. Representative: Blue Yonder Cognitive Solutions, Kinaxis Maestro AI, o9 AI Agents, Manhattan Active Agents, project44 AI Agents, FourKites Digital Workforce. (o9 is the most Native-leaning, AI-forward from a 2009 founding; kept Embedded on Test 1 because the knowledge-graph/optimization planning core delivers value without generative AI. SAP IBP held pending GA of its planning-specific agentic AI; e2open held on watchlist post-WiseTech integration.)

Process Automation & BPM (10 products). AI capability embedded in established process-automation, BPM, RPA, and workflow-orchestration platforms that predate the AI and run without it. Boundary: in scope are AI copilots, agents, and agentic orchestration inside an incumbent automation/BPM platform; no Native counterpart cell exists today (AI-first agent build-and-run platforms sit under the Native Agent Platforms & Frameworks cell). Representative: Pega GenAI, Appian AI Agents, Automation Anywhere Agentic Process Automation, SS&C Blue Prism WorkHQ, Nintex AI, Microsoft Power Automate, UiPath Agentic Automation, Camunda Agentic Orchestration, Workato AIRO, MuleSoft for Agentforce. (Workato and MuleSoft are iPaaS/integration platforms placed here as the nearest host category since the Atlas has no dedicated iPaaS cell; a dedicated Integration / iPaaS cell is a future split candidate.) (UiPath and Automation Anywhere are the deepest agentic, Level 3; all remain Embedded on Test 1 and Test 3, since the deterministic RPA/BPM cores predate and run without the AI.)

Development & DevOps Platforms (3 products). AI capability embedded in established development and DevOps platforms. Boundary: in scope are AI features inside a developer platform; the Native counterpart is Software Development & Coding. Representative: GitHub Copilot, Tricentis Aida (agentic quality engineering), Heroku AI (Managed Inference and Agents).

Industry Clouds (5 products). AI capability embedded in established industry-specific suites and platforms. Boundary: in scope are vertical incumbents whose core platform predates and survives without the AI; the Native counterparts are the matching Vertical Application cells. Representative: Lexis+ AI (Protégé, legal), Innovaccer (healthcare), Cognite Atlas AI (industrial), Ellucian Student (higher education), Oracle Health Clinical AI Agent.

5. Category census (September 2026, after the Salesforce reassessment)

Panel totals are read directly from the Atlas base rather than summed by hand, so they are the authoritative current figures, as of 2026-09-05. On-map products (flagged Included on the map, across all agentic levels): 387 — 214 Native, 173 Embedded — across 46 category cells, 43 populated. Catalogued products including entries tracked but not yet on the map: 406. Vendors: 334. Sources: 787 (at least two per product).

The only unpopulated cells in the entire Atlas are the three Native orchestration cells (Model & Agent Gateways, Interoperability & Protocols, Agent Memory & Context). The Embedded panel is fully populated across all 17 host categories. Embedded depth is concentrated in Marketing Technology (22, now tied for the deepest single cell), Security Platforms (22), Analytics & BI Platforms (22), and HCM & Talent Suites (16).

Note on reconciliation: the per-cell counts shown in Sections 3 and 4 are updated at each sprint and reflect the cells touched in that sprint; between full recounts they can trail the live panel totals slightly as untracked additions land in the base. From v0.10 the Section 5 panel totals are reconciled to the live base, which is why they step up from the v0.9 hand-summed figures.

Part II. Business Function and Use Case

This lens answers the buyer's question: I run a function, what can AI do for me. It is organized in two levels. Business function is the primary axis, who has the problem. Use case is what the product does and the business problem it solves. A product is tagged with one to four primary use cases, and its function follows from them; a large function such as Information Technology also carries secondary sub-functions. This lens is orthogonal to Part I, so a product carries both a structural placement and one or more use cases.

The 13 functions

Each function below carries a one-line definition, the typical buyer, a boundary note where a merge or overlap needs a rule, and a curated use-case list. In each use case the parenthetical is the business problem it solves. Example Atlas products confirm coverage.

1. Go-to-Market

Sales, marketing, sales ops, and marketing ops. Buyer: CRO, CMO, VP Sales, VP Marketing, RevOps.

  • Lead research and enrichment (reps waste hours on manual prospecting)

  • Outbound outreach and AI SDR (scaling personalized pipeline generation)

  • Content and creative generation (content production bottleneck and cost)

  • Campaign and journey orchestration (manual, slow campaign builds)

  • Audience segmentation and personalization (generic, one-size messaging)

  • Conversation intelligence and call capture (insight and CRM data lost after calls)

  • Deal coaching and next-best-action (inconsistent rep performance)

  • Revenue forecasting and pipeline analytics (unreliable pipeline calls)

  • SEO and AEO and web optimization (weak visibility in search and AI answers)

  • RevOps data hygiene and enrichment (dirty CRM and GTM data)

Atlas: Clay, Rox, 11x, Qualified, 6sense; Agentforce, Salesforce Marketing Cloud, Adobe Marketo Engage, Adobe Journey Optimizer, GenStudio, Braze, Klaviyo; Jasper, Synthesia, Adobe Firefly, Webflow.

2. Customer Service and Support

Customer service, support, and contact center. Buyer: VP Customer Experience, Head of Support, Contact Center leader.

  • Self-service deflection and customer-facing agents (ticket volume overwhelms teams)

  • Agent assist and copilots (slow, inconsistent responses)

  • Ticket triage, routing, and summarization (misrouted and slow handling)

  • Voice and contact-center automation (staffing cost and wait times)

  • Knowledge management and answer generation (answers buried or stale)

  • Quality monitoring and agent coaching (inconsistent quality at scale)

  • Proactive support and churn signals (support is reactive only)

Atlas: Sierra, Decagon, Intercom Fin, Ada, Cresta, PolyAI; NiCE CXone Mpower, Genesys Cloud AI, Five9 Genius AI, Talkdesk, Twilio CustomerAI, Sprinklr AI+.

3. Finance and Accounting

Controllership, FP&A, treasury, tax, and transactional finance. Buyer: CFO, Controller, VP Finance.

  • Financial close and reconciliation (slow, error-prone close)

  • Accounts payable and receivable automation (manual invoice and collections work)

  • FP&A, budgeting, and forecasting (spreadsheet-bound planning)

  • Expense management and audit (leakage and compliance risk)

  • Financial research and analysis (analyst time lost gathering data)

  • Treasury and cash management (limited cash visibility)

  • Revenue recognition and billing (complex, error-prone rev rec)

  • Tax and statutory reporting (manual, deadline-driven risk)

Atlas: Basis, Rillet, Campfire, Numeric; SAP Joule for Finance, Oracle Fusion ERP AI, NetSuite AI, Microsoft Dynamics 365 Finance Copilot, BlackLine, Workiva AI Agents, Intuit Assist.

4. Procurement and Sourcing

Source-to-pay, spend management, and supplier management. Buyer: CPO, Head of Procurement and Sourcing.

  • Sourcing and RFx automation (slow supplier selection)

  • Contract lifecycle management (bottlenecked contracting)

  • Spend analysis and optimization (no visibility into spend)

  • Supplier discovery, risk, and performance (supply exposure)

  • Intake-to-procure and guided buying (maverick spend, slow requests)

Atlas: Coupa Navi, SAP Ariba, GEP, JAGGAER, Ivalua, Zip, LightSource, Workday CLM.

5. Supply Chain and Logistics

Planning, inventory, and the movement of goods. Buyer: Chief Supply Chain Officer, VP Operations, VP Logistics.

  • Demand and supply planning (forecast error)

  • Inventory optimization (stockouts versus overstock)

  • Logistics and real-time transportation visibility (blind spots in transit)

  • Warehouse and fulfillment optimization (labor and throughput)

  • Supply-chain risk and network design (disruption exposure)

  • Control-tower decision intelligence (slow cross-network decisions)

Atlas: Blue Yonder, Kinaxis, o9, Manhattan, project44, FourKites; Aera, Altana, Interos.

6. Production and Service Delivery

Manufacturing, product, and service delivery: the making and delivering side of operations. Buyer: COO, VP Manufacturing, Plant leader, Head of Service Delivery. Boundary: the planning-and-moving side (planning, inventory, logistics) sits in Supply Chain and Logistics; Production owns execution, quality, uptime, and delivery.

  • Manufacturing process optimization and scheduling (throughput and downtime)

  • Quality inspection and defect detection (defect escapes and scrap)

  • Predictive maintenance and asset performance (unplanned downtime)

  • Field service optimization (technician efficiency and first-time-fix)

  • Service and project delivery automation (delivery cost and consistency)

  • Worker safety and EHS monitoring (incidents and compliance risk)

Atlas: Augury, Instrumental, Carbon Robotics, Everseen; Cognite Atlas AI.

7. Human Resources

Talent acquisition, HR operations, payroll, and people analytics. Buyer: CHRO, VP People, VP Talent.

  • Recruiting and sourcing (time-to-fill)

  • Screening and interviewing (volume and bias)

  • Onboarding (slow ramp)

  • Employee self-service and HR helpdesk (repetitive queries)

  • Talent development and skills intelligence (skills gaps and retention)

  • Payroll and workforce anomaly detection (costly errors)

  • Workforce planning and analytics (reactive people decisions)

Atlas: Paradox, Juicebox, Eightfold, Metaview; Workday Illuminate, isolved, Rippling, SAP SuccessFactors, ADP Assist.

8. Research and Development

Building the products the company sells: software and physical products. Software engineering, product management, product strategy, and product and UX design. Buyer: CTO, Chief Product Officer, VP Engineering, VP Product, VP Design. Boundary: R&D builds the company's own products; running the company's internal technology is IT. A coding tool is tagged where its buyer sits and can carry both when it genuinely serves product engineering and internal platform teams.

  • Code generation (developer throughput)

  • Code review and test generation (quality and coverage)

  • Agentic software delivery (backlog velocity)

  • Quality engineering and test automation (release risk)

  • Product management and requirements (roadmap and spec time)

  • Product and UX design and prototyping (design-to-build lag)

  • Engineering knowledge and code search (ramp time and context switching)

  • Physical product and R&D design, simulation, and discovery (long design and discovery cycles, including CAD, materials, and life-sciences discovery)

Atlas: Cursor, Claude Code, Devin, Replit; GitHub Copilot, Tricentis Aida, Tabnine; Recursion OS (discovery).

9. Information Technology (IT)

The company's internal technology estate: IT and service management, security, IT operations and AIOps, internal DevOps and platform engineering, datacenter and cloud infrastructure, and networking and communications. Buyer: CIO, CISO, VP Infrastructure, Head of SecOps. Boundary: internal tech estate, distinct from product engineering (R&D) and from enterprise risk as a strategy topic (Strategy and Governance). This is the broadest function; for navigation it is best shown with sub-groups.

  • IT service management and helpdesk (ticket load and slow resolution)

  • IT operations, AIOps, and observability (alert noise and MTTR)

  • Security operations and threat detection (alert fatigue and analyst shortage)

  • Vulnerability and posture management (exposure)

  • AI and agent security and governance (new attack surface)

  • Identity and access management (access risk)

  • DevOps, platform engineering, and incident response (deployment risk and reliability)

  • Infrastructure, datacenter, and cloud operations (cost and complexity)

  • Networking and communications (network reliability and comms)

Atlas: ServiceNow AI Agents, Moveworks; Microsoft Sentinel, CrowdStrike, Palo Alto, SentinelOne; HiddenLayer, Zenity, Noma; Datadog, Nutanix Agentic AI.

10. Legal and Compliance

Legal work and regulatory compliance. Buyer: General Counsel, Chief Compliance Officer, Head of Legal Ops. Boundary: legal and regulatory only; enterprise and board governance and ESG sit in Strategy and Governance; IT and security compliance sits in IT.

  • Contract review and CLM (legal review bottleneck)

  • Legal research (research time)

  • Regulatory compliance monitoring (regulatory exposure)

  • E-discovery and investigations (document volume)

  • Legal knowledge and matter management (matter cost and visibility)

Atlas: Harvey, Legora, CoCounsel, Robin AI, Luminance; Lexis+ AI, Workday CLM.

11. Data, Analytics and BI

Horizontal data and analytics platforms and the work of turning data into insight. Buyer: Chief Data Officer, VP Data, VP Analytics. Boundary: horizontal data and BI; domain-specific analytics (revenue forecasting, supply planning) live with their function.

  • Self-service and natural-language analytics (analyst backlog)

  • Data preparation, integration, and pipelines (dirty, model-unready data)

  • Data quality, catalog, and governance (untrusted data)

  • Anomaly detection and automated insight (missed signals)

  • Embedded and product analytics (decisions made without data in the workflow)

Atlas: Hex, Julius; Power BI Copilot, Tableau, ThoughtSpot, Adobe CJA; Databricks Mosaic AI, Snowflake Cortex.

12. Business Strategy, Planning and Governance

Corporate and enterprise level: strategy, planning, corporate development, enterprise risk, and governance. Buyer: CEO, COO, Chief Strategy Officer, Corp Dev, Board, Enterprise Risk. Boundary: corporate scope. FP&A is Finance, legal and regulatory is Legal, security is IT.

  • Market and competitive intelligence (slow, shallow market insight)

  • Strategic and scenario planning (reactive strategy)

  • Enterprise and integrated business planning (siloed plans; EPM overlaps Finance)

  • M&A and corporate development research and diligence (deal research time)

  • Enterprise risk management (blind spots in enterprise risk)

  • Board, governance, and ESG reporting (governance and disclosure burden)

  • Executive decision support and knowledge (slow executive decisions)

Atlas: Govini Ark (public-sector strategy); Anaplan (planning, shared with Finance); horizontal research tools applied to strategy.

13. Workplace Productivity and Collaboration

Horizontal knowledge work that cuts across every function. Buyer: typically IT or HR sponsored, used by everyone.

  • Meeting capture and summarization (lost context)

  • Document drafting and summarization (writing time)

  • Enterprise search (answers scattered across tools)

  • Personal and team assistants and copilots (task overhead)

  • Knowledge management and internal wiki (tribal knowledge)

  • Translation and transcription (language and access barriers)

Atlas: Granola, Otter, Glean, Perplexity Enterprise; Microsoft 365 Copilot, Notion AI, Zoom AI Companion, Box AI, Egnyte, Asana.

Cross-functional and Horizontal foundations

A deliberate catch-all for products that power every function and belong to none: foundation models, agent build-and-run platforms, inference and compute, vector and data infrastructure, evals and observability, and AI governance platforms. Kept out of the functional views so horizontal products do not bloat every function.

Atlas: Claude for Enterprise, ChatGPT Enterprise, Google Gemini; LangGraph, CrewAI, Bedrock AgentCore, StackAI; CoreWeave, Together AI, Pinecone; Arize, LangSmith; Credo AI, OneTrust.

Cross-cutting tagging rules

  • Tag at the use-case level; function rolls up. A product links to one to four primary use cases. Its function(s) derive from those links.

  • Cap primary tags per product at roughly four. The question is what a product is mainly for, not every conceivable use. Horizontal products go to the Cross-functional bucket, not to ten functions.

  • A use case has one primary function, and may surface under a second only when it genuinely straddles (contract lifecycle management under Procurement and Legal; EPM under Finance and Strategy; coding assistants under R&D and IT).

  • Curated additions only. A new use case earns a row the way a new category earns a cell: with a written business problem and a logged reason.

Maintenance rules

Boundaries are enforced by the methodology, not by category names. When a product could fit two cells, place it by its primary value and record the alternative in the rationale.

A cell that grows past roughly a dozen well-differentiated products is a candidate to split; a cell that stays at one or two members is a candidate to merge or to fold into a broader cell, unless it is deliberately reserved as roadmap. The Security Platforms cell (22), the Analytics & BI Platforms cell (22), and the Marketing Technology cell (22) are the clearest split candidates. Security Platforms could split along SIEM/SecOps versus endpoint/XDR versus other embedded security modules; Analytics & BI along hyperscaler/enterprise BI, independent BI, and product/embedded analytics; Marketing Technology could shed a dedicated Creative & Design Tools cell (pairing with the Native Content & Creative panel) that would take Adobe Creative Cloud's embedded AI and future Canva/Figma entries. The HCM & Talent Suites (16), Agent Platforms & Frameworks (15), ERP & Finance Suites (14), and Labeling & Synthetic Data (12) cells are also on split-watch. The single-member Embedded cells (Privacy & GRC Platforms, Development & DevOps Platforms) are merge-watch only if they fail to grow. The Other Verticals holding cell should shed industries into dedicated cells once any one reaches critical mass. A dedicated Embedded Content Management cell (currently Box and Adobe Acrobat Studio under Collaboration & Productivity Suites) is a future candidate if that segment deepens.

Every category change (add, split, merge, rename, or redefinition) is recorded in the edition delta with a reason, so the taxonomy stays auditable across editions.

Version history

  • v0.13 (September 2026). Consolidated into a single taxonomy document with two parts, Part I (Structure: Native and Embedded) and Part II (Business Function and Use Case), under one version and one version history. The function/use-case (demand-side) lens, previously a separate document, was folded in: 13 primary functions plus a Cross-functional / Horizontal Foundations bucket, a curated 94-use-case library with the business problem each solves, and a primary/secondary function hierarchy (Information Technology seeded as the first sub-function group). The lens is live in Airtable (Functions and Use Cases tables linked to Products, with a function rollup); a three-function pilot (Go-to-Market, Customer Service and Support, Finance and Accounting) tagged 81 products and is surfaced through a buyer-facing "Browse by Function" page. Table and field IDs are recorded in the database schema doc; the demand-lens build steps are logged in the sprint log.

  • v0.12 (September 2026). Salesforce reassessment (6 products, 0 new vendors; all reused the Salesforce vendor record, consistent with the Tableau precedent). Salesforce went from four mapped products to ten. Added Slack (Slack AI / Agentforce in Slack) (Embedded, Collaboration & Productivity Suites, Level 2); MuleSoft for Agentforce (Embedded, Process Automation & BPM as nearest host, Level 2); Informatica IDMC (CLAIRE AI / CLAIRE Agents) (Embedded, Data Platforms & Clouds, Level 2 — Salesforce completed the Informatica acquisition Nov 18, 2025); Salesforce Data Cloud (Embedded, Data Platforms & Clouds, Level 1, the Agentforce grounding/vector layer); Salesforce Agentforce Commerce / Commerce Cloud (Embedded, CRM & Customer Experience Platforms as nearest host, Level 2, with a future Commerce cell flagged); and Heroku AI (Managed Inference and Agents) (Embedded, Development & DevOps Platforms, Level 2). Cell counts updated to live: CRM & Customer Experience Platforms 11→12, Collaboration & Productivity Suites 7→9, Data Platforms & Clouds 9→12, Process Automation & BPM 9→10, Development & DevOps Platforms 2→3 (some of these absorb prior untracked drift). A future Integration / iPaaS cell (Workato, MuleSoft) and Commerce cell were flagged. Census re-read live: on-map 387 (214 Native / 173 Embedded), 406 catalogued, 334 vendors, 787 sources. Also produced a public-facing merged taxonomy (category + function/use-case lenses) for the website: claude/atlas-taxonomy-public.md.

  • v0.11 (August 2026). Approved-backlog application sprint (Aug 25): applied the 15 Review Queue rows Michael had approved. Nine new products (with 8 new vendors): Tricentis Aida (Embedded, Development & DevOps Platforms), LightSource AI Workforce (Embedded, Procurement), Workiva AI Agents (Embedded, ERP & Finance Suites), Fisent BizAI (Native, Financial Services), Drata AI Agent Governance (Embedded, Privacy & GRC Platforms, Level 1), Everseen Vision AI Platform (Native, Other Verticals), Egnyte AI Agents (Embedded, Collaboration & Productivity Suites), Workato AIRO (Embedded, Process Automation & BPM, iPaaS placed as nearest host), and xAI Grok Bot (Native, Agent Platforms & Frameworks). Two agentic upgrades to existing entries: WRITER AI HQ (to Level 3, Palmyra X6 / 8-hour unattended workflows) and isolved People Cloud (The Guardian payroll agent, Level 2). Three acquisition/ownership updates on existing products (classification unchanged, vendor ownership set to Subsidiary): OpenRouter (Stripe, pending close), Cursor/Anysphere (SpaceX, completed), Tabnine (Tricentis, kept standalone Native). One "other" update: Intuit Assist / Intuit Enterprise Suite (Intuit Intelligence expansion, no reclassification). Two more products, promoted from the web-submission form and approved by Michael the same day, were also applied: StackAI (Native, Agent Platforms & Frameworks, Level 2) and Asana Agentic Work Management (Embedded, Collaboration & Productivity Suites, Level 2), each with a new vendor record. Cell counts updated: Development & DevOps Platforms 1→2, Procurement 6→7, ERP & Finance Suites 13→14, Financial Services 8→9, Privacy & GRC Platforms 1→2, Other Verticals 5→6, Collaboration & Productivity Suites 6→8, Process Automation & BPM 8→9, Agent Platforms & Frameworks 13→15. Section 5 census re-read live from the base (on-map 378: 214 Native / 164 Embedded; 394 catalogued; 332 vendors; 763 sources), which also absorbs other additions applied to the base since v0.10 (the base is actively maintained between sprints). Follow-on (Aug 30): added Workday Contract Lifecycle Management (CLM), powered by Evisort AI (Embedded, Procurement as nearest host, Level 2), reusing the Workday vendor record; a dedicated Contract Management / CLM cell was flagged, and Workday's separate Contract Intelligence product (also Evisort-based) noted as a candidate. Census stepped to on-map 379 (214 Native / 165 Embedded), 395 catalogued, 332 vendors, 765 sources; Procurement cell 7→8. Follow-on (Sept 3): added Webflow (Embedded, Marketing Technology as nearest host, Level 1; new vendor) after an add/hold check — its visual web-dev/CMS core predates the AI, GA AI features clear the bar, and the agentic Source by Webflow is in limited preview (Level 2 upgrade path). Ploy was re-checked the same day and held on Horizon (Native, still the June 2026 $27M seed, below the enterprise bar). Marketing Technology cell 21→22 (now tied deepest with Security Platforms and Analytics & BI at 22); a future Web Experience / DXP cell flagged. Census stepped to on-map 381 (214 Native / 167 Embedded), 400 catalogued, 334 vendors, 775 sources. Wording of the census note corrected: on-map = flagged Included in Map across all agentic levels, not level 2+ only.

  • v0.10 (August 2026). Adobe deep-dive sprint (7 products, 0 new vendors; the Adobe vendor record was reused). Adobe went from one mapped product to eight. Added Adobe Firefly (Native, Content & Creative — the model family is the product); Adobe Experience Platform Agent Orchestrator (Embedded, Marketing Technology, Level 3), Adobe Journey Optimizer, and Adobe GenStudio for Performance Marketing (Embedded, Marketing Technology); Adobe Creative Cloud embedded AI / Creative Agent (Embedded, routed into Marketing Technology as the nearest host category, with a future Creative & Design Tools split flagged); Adobe Customer Journey Analytics (Embedded, Analytics & BI Platforms); and Adobe Acrobat Studio AI Assistant (Embedded, Collaboration & Productivity Suites). The existing Marketo Engage record was renamed to drop its AEP-Orchestrator parenthetical now that Agent Orchestrator is its own entry. Cell counts updated: Content & Creative 7→8, Marketing Technology 17→21, Analytics & BI Platforms 21→22, Collaboration & Productivity Suites 5→6. From this edition the Section 5 panel totals are reconciled to the live base (on-map 354: 202 Native / 152 Embedded; 308 vendors; 699 sources), which also absorbs earlier untracked additions (including the C3 AI and Palantir entries) logged to the base and the public site but not previously itemized in this document.

  • v0.9 (August 2026). Two sprints folded in. Conference-host classification sprint (8 products, 8 vendors): added Zoom AI Companion and Box AI (Collaboration & Productivity Suites), Sprinklr AI+ (CRM & CX), Snowflake Cortex AI (Data Platforms & Clouds), Twilio CustomerAI (Contact Center & CCaaS), Coupa Navi (Procurement), and Aera Decision Cloud (Native Manufacturing & Supply Chain); Ploy held on the Sales & Marketing native watchlist (out of stealth June 2026, below the enterprise bar). Supply-chain and process-automation sprint (18 products, 15 vendors): split the former "Procurement & Supply Chain" cell into Procurement (Coupa, SAP Ariba, GEP, JAGGAER, Ivalua, Zip) and a new Supply Chain & Logistics cell (Blue Yonder, Kinaxis, o9, Manhattan, project44, FourKites); added a new Process Automation & BPM cell (Appian, Automation Anywhere, SS&C Blue Prism, Nintex, Microsoft Power Automate, UiPath, Camunda) and moved Pega GenAI into it from CRM & CX. SAP IBP and e2open held on watchlist. With these, the Embedded panel is fully populated (17 of 17 host categories). Census refreshed to 345 products (194 Native, 151 Embedded), 298 vendors, 659 sources, 43 of 46 cells populated. New pairing added to Section 2 (Manufacturing & Supply Chain Native with Procurement and Supply Chain & Logistics Embedded). Note: a public "emerging / pre-inclusion" vendor surface (working name pending) is planned as a Q4 data-product initiative, drawing on the watchlist candidates recorded across sprints.

  • v0.8 (August 2026). Security endpoint follow-on: Security Platforms (Embedded) deepened from 16 to 22 by adding Google SecOps (Gemini in Security) and the endpoint/XDR field (SentinelOne Purple AI, Trend Vision One / Trend Companion, Sophos AI Assistant, Bitdefender GravityZone / CLAIRE, Tanium AI). Google (Alphabet) and SentinelOne vendor records reused. Microsoft Defender XDR held (same Security Copilot already mapped via Sentinel); Trellix, Cybereason, Broadcom/Symantec CBX, WithSecure, ESET held on watchlist. Census refreshed to 319 products across 40 populated cells; Security Platforms (22) is now the deepest single cell in either panel.

  • v0.7 (August 2026). Security SIEM sprint: Security Platforms (Embedded) deepened from 2 to 16 using the G2 SIEM Grid as the market-presence authority. Added the enterprise SIEM/SecOps field (Microsoft Sentinel + Security Copilot, CrowdStrike Falcon Charlotte AI, Splunk Enterprise Security, Palo Alto Cortex XSIAM, IBM QRadar watsonx, Securonix EON, Exabeam Nova, Rapid7 Incident Command, Elastic Security, Datadog Cloud SIEM, Sumo Logic, Fortinet FortiSIEM, Check Point Infinity, OpenText ArcSight). Kept the Native/Embedded split clean: Prisma AIRS (Palo Alto) and Lakera (Check Point) remain in the Native AI Security cell as distinct products. Mid-market/MSP SIEMs held on watchlist; Adlumin and NetWitness skipped (no named generative/agentic AI). Census refreshed to 313 products across 40 populated cells; AI Security / Security Platforms pairing added to Section 2.

  • v0.6 (August 2026). HCM sprint: HCM & Talent Suites (Embedded) deepened from 2 to 16 using the G2 HCM Software Grid as the market-presence authority. Added enterprise suites (SAP SuccessFactors Joule, ADP Workforce Now, UKG Pro Bryte AI, Dayforce Co-Pilot, Infor HCM, Darwinbox) and modern mid-market platforms (Rippling, Paycom, Paylocity, HiBob, Deel, Personio, isolved, Paychex). Workday Illuminate and Oracle Fusion HCM already held the cell and were not duplicated; BambooHR, Gusto, and Paycor held on watchlist; Namely skipped. Census refreshed to 299 products across 40 populated cells; HR & Talent / HCM & Talent Suites pairing added to Section 2.

  • v0.5 (August 2026). Marketing technology sprint: Marketing Technology (Embedded) populated from Roadmap to 17, using the G2 Marketing Automation grid as the market-presence authority. Members span marketing-cloud/suite incumbents and the customer-engagement/email field. HubSpot Marketing Hub held as a duplicate of Breeze (CRM & CX); Zoho Marketing Automation and Marigold noted but not added; four AI-native martech tools held on the Sales & Marketing native watchlist. Census refreshed to 285 products across 40 populated cells.

  • v0.4 (August 2026). Analytics expansion: Analytics & BI Platforms (Embedded) deepened from 3 to 21 with the legacy and modern BI field. Also added Oracle AI Agents for Fusion Cloud HCM, Oracle Health Clinical AI Agent, and Tableau AI. Census refreshed to 268 products; Analytics & BI flagged as the top split candidate.

  • v0.3 (August 2026). Sprint 6 update: Embedded ERP & Finance Suites cell populated with 13 incumbents; census refreshed to 246 products across 39 populated cells. Finance & Operations / ERP & Finance Suites pairing added to Section 2.

  • v0.2 (August 2026). Sprint 5 update: all six Native Vertical Application cells populated and the Embedded Industry Clouds cell seeded with its first four members; census refreshed to 233 products across 38 populated cells. Vertical/Industry Clouds pairing added to Section 2.

  • v0.1 (August 2026). Initial taxonomy: 44 cells across both panels, boundaries and representative products for the populated cells, roadmap cells marked. Companion to the Native & Embedded Classification Methodology v0.1.