Skip to main content

AI Product Taxonomy

Executive summary

The enterprise AI market describes a wide set of solution classes — from corporate LLM access to optimization and digital twins. This page captures 18 classes as a reference / research taxonomy: it helps classify business requests and decide when a class deserves its own AI product.

The AI operating model keeps the real catalog at no more than 10 delivery tracks — the routes through which AI initiatives move, get owners, and pass stage gates. Other classes serve as reference material, industry slots, or cross-cutting layers.

For the operational catalog, see AI products.


How to read the taxonomy

This page uses different levels of abstraction:

LevelWhat it is
Model / platformReusable technical capability (LLM gateway, RAG, ML platform)
AI productManaged solution with an owner, SLA, and usage rules
Capability / layerFunction inside a product (guardrails, eval, data readiness)
AI initiativeA specific use case in the business funnel
Delivery trackImplementation route in the delivery funnel

The key question: does this need a separate AI product and delivery track, or is a capability inside an existing track enough?


Full taxonomy: 18 classes

Compact reference table. Tier reflects typical enterprise demand maturity — not mandatory standalone tracks.

#ClassTierWhat it isVerbsInclude whenDo not create separately when
1LLM Self-Service & Model GatewayCoreGoverned access to approved foundation models via UI/API: chat, playground, routing, templates, cost controls, governancegenerate, summarize, classify, extract, rewrite, translate, reason, draft3+ functions need GenAI or multiple initiatives need reusable model accessM365/Google Copilot covers the need with adequate controls, or it is only a feature inside one app
2Enterprise Knowledge Search / RAG / Q&ACoreGrounded search and Q&A over enterprise data with citations, permission-aware retrieval, and traceabilityfind, search, answer, summarize, compare, retrieveSame Q&A/search pattern appears in 3+ functions or 5+ knowledge domainsOne-document chatbot, data is not governed, or enterprise search already solves it
3Agentic Workflow / Routine AutomationCoreOrchestrates LLM agents, deterministic workflows, tools, APIs, humans, and approvals for multi-step tasksautomate, orchestrate, collect, decide, act, monitor, escalateHigh-volume routine work repeats across functions; APIs or clear approval gates existOne-off expert work, process is undefined, no system access, or RPA/iPaaS already solves it
4Document Intelligence / IDPCoreDocument classification, splitting, OCR, extraction, validation, and routing into business systemsextract, classify, process, split, compare, validate, route3+ processes or document types need repeatable extraction and validationLow volume, data available via API, or it is a one-off form
5Meeting & Conversation IntelligenceAdvancedCaptures, transcribes, and summarizes meetings/calls; extracts decisions and action itemstranscribe, summarize, extract actions, coachMeetings/calls are high-volume work artifacts; follow-up is a recurring painLegal/culture prevents recording, or native M365/Zoom capability is sufficient
6Coding Agent / Software Engineering CopilotCoreIDE/CLI/cloud agents for repo understanding, code generation, refactoring, testing, debugging, PR assistancewrite code, explain, refactor, test, debug, document, migrateSoftware delivery is material; enough developers and repos to justify governanceFew developers, or a centrally procured IDE add-on with no portfolio management need
7ML Platform / Model Factory / Decision AICorePlatform for classical ML, scoring, forecasting, anomaly detection, recommendations, and model lifecyclepredict, score, classify, forecast, recommend, detect anomalyMultiple predictive/decision use cases recur; historical data existsProblem is search/generation; no labels, business owner, or action path
8BI / Analytics Copilot / Decision IntelligenceAdvancedNatural-language and agentic analytics over governed semantic data, metrics, and dashboardsask data, analyze, compare, explain variance, monitor KPIsKPI/analytics questions repeat; governed semantic data existsData is not trusted, or each analysis is bespoke consulting
9Customer & Employee Service AgentAdvancedAI agents for support/service desks: answer, triage, resolve, tickets, escalationanswer, classify, route, resolve, automate, escalateHigh support volume; KB and actions are reusableLow-volume expert advice or no KB/action integration
10Voice / Speech AI & Contact CenterOptionalReal-time speech, voice agents, agent assist, call summarization, QA, analyticstranscribe, summarize, answer by voice, route, coachVoice is a core channel or contact-center volume is highMeeting transcription is enough, or legal consent is infeasible
11AI Governance / Portfolio Management / Control TowerCoreSystem of record for AI initiatives, models, agents, risks, approvals, ownership, value, compliancegovern, inventory, approve, prioritize, monitor, reportLarge or regulated company with multiple AI products/agentsFewer than ~10 AI uses; start with lightweight GRC/PMO workflow
12AI Security / Guardrails / Red TeamingCoreSecurity controls and testing for AI apps, LLMs, agents, prompts, data, model supply chainprotect, detect, block, redact, monitor, test, respondAI touches sensitive data, external users, or tool/actionsLow-risk prototype; embed minimum controls inside the LLM platform
13AI Evaluation, Observability & LLMOpsCoreTesting, tracing, monitoring, and improving LLM apps, RAG, agents, prompts, costs, qualityevaluate, monitor, debug, trace, compare, optimize3+ production LLM/RAG/agent appsVendor black-box without instrumentation; use vendor monitoring instead
14Data & Knowledge Readiness / AI Data ProductCoreReusable layer: catalog, access, lineage, quality, classification, semantic definitions, knowledge readinessconnect, govern, classify, enrich, curate, retrieve, reuseData readiness is a recurring bottleneck across AI initiativesMature enterprise data platform already owns it; treat as capability, not duplicate
15Multimodal Content Generation / Brand CreativeOptionalControlled generation/editing of text, image, video, audio, design with brand/legal controlsgenerate, edit, localize, personalize, version, reviewHigh creative/localization volume across business unitsOccasional design work, or agency/DAM workflow already covers it
16Computer Vision / Visual AI / Edge AIOptionalAnalyzes images/video/physical environments: detection, inspection, safety, process visibilitydetect, classify, count, inspect, monitor, alertPhysical operations have repeatable visual problemsOffice knowledge work, or sensor/API data is enough
17Research / Expert Workbench / High-Stakes AnalysisAdvancedExpert workbench for corpus analysis, source comparison, evidence-backed memos, high-stakes decisionsresearch, compare, diligence, synthesize, citeHigh-value expert analysis repeats across functionsGeneral RAG or BI copilot already covers the task
18Optimization / Simulation / Digital Twin / Prescriptive AIOptionalAI/OR/simulation for optimal plans under constraints and scenario simulationoptimize, simulate, recommend, allocate, schedule, prescribeDecisions are high-value, repeated, codifiable with constraints/dataAd hoc decisions, unreliable data, or planning system already optimizes well

Mapping to the operational catalog

The operational catalog stays at no more than 10 delivery tracks. Industry slots A and B are reserved for recurring company-specific demand (analytics, voice, creative, CV, research, optimization — depending on context).

The 10 operational tracks:

#Delivery track
1Corporate LLM
2RAG / Knowledge Assistant
3ML platform
4Code agent
5Automation AI (agents + orchestration)
6Document Intelligence
7Meeting intelligence
8Service Agent
9Industry slot A
10Industry slot B

18 classes → default operational handling:

Research classDefault operational handling
LLM Self-Service & Model GatewayTrack 1: Corporate LLM
Enterprise Knowledge Search / RAG / Q&ATrack 2: RAG / Knowledge Assistant
ML Platform / Model Factory / Decision AITrack 3: ML platform
Coding Agent / Software Engineering CopilotTrack 4: Code agent
Agentic Workflow / Routine AutomationTrack 5: Automation AI
Document Intelligence / IDPTrack 6: Document Intelligence
Meeting & Conversation IntelligenceTrack 7: Meeting intelligence
Customer & Employee Service AgentTrack 8: Service Agent
BI / Analytics CopilotTrack 9/10 if analytics is a repeated strategic demand; otherwise applied capability over the data/BI stack
Voice / Speech AITrack 9/10 for contact-center-heavy companies; otherwise a channel inside Service Agent
Multimodal Content FactoryTrack 9/10 for marketing/content-heavy companies
Computer Vision / Edge AITrack 9/10 for manufacturing, retail, logistics, security, field operations
Research / Expert WorkbenchTrack 9/10 for legal, risk, investment, regulatory contexts; otherwise advanced RAG pattern
Optimization / Digital TwinTrack 9/10 for planning-heavy industries; often an ML/OR extension
AI Governance / Portfolio ManagementCross-cutting governance layer — not a separate track by default
AI Security / Guardrails / Red TeamingCross-cutting security layer — not a separate track by default
AI Evaluation, Observability & LLMOpsCross-cutting quality/release layer — not a separate track by default
Data & Knowledge ReadinessCross-cutting data/knowledge layer — not a separate track by default

Governance, AI security, eval/observability, and data readiness are research classes and cross-cutting layers by default. A separate delivery track is justified only with a clear owner and repeated company demand.


Name synonyms

The market uses different names for the same classes. Synonyms help map market language to operational tracks without catalog sprawl:

Operational trackMarket synonyms
RAG / Knowledge AssistantEnterprise Knowledge Search, grounded Q&A, knowledge assistant
Automation AIAgentic Workflow, agentic automation, routine automation platform
Code agentCoding Agent, Software Engineering Copilot, dev copilot

Sources and market anchors

The links below are market anchors: they help orient in the current landscape but are not timeless methodology facts. Forecasts and vendor positioning change.

Product examples (to clarify classes only — not recommendations):