The AI Pattern Vendor Landscape Map

There are hundreds of AI vendors. Procurement processes stretch to months not because buyers are slow, but because most operators don't have a framework for categorizing what they're evaluating. Every vendor claims to be "AI-powered" and "enterprise-ready." Every demo looks capable. Every pricing page obscures what you're actually buying.

Pattern thinking cuts through vendor marketing instantly. Instead of asking "what features does this vendor have," ask: "What pattern does this vendor serve?" The answer tells you which business problem the product is solving, how mature the product category is, who should own the procurement decision internally, and how to evaluate whether the vendor's version of the pattern fits your use case.

This is an orientation map, not a buying guide. It tells you what product categories exist for each pattern and how mature each category is. It doesn't tell you which vendor to buy. That decision requires your specific data model, your integration requirements, and your security review. For independent vendor rankings within specific pattern categories, the Gartner Magic Quadrant for Conversational AI Platforms and the Forrester Wave for AI Decisioning Platforms, Q2 2025 are the two most referenced analyst evaluations in enterprise AI procurement.

How to use this map

For each pattern, you'll find: the product category name (what vendors call themselves in this market), maturity rating (high/moderate/emerging), the typical internal buyer persona, and a description of how the vendor landscape is structured.

Vendor names appear only as category examples. They're not rankings or endorsements.

Key Facts: AI Vendor Market Scale

• The enterprise AI market reached $114.87 billion in 2026 and is projected to grow at 18.9% CAGR through 2031, reaching $273 billion. (Mordor Intelligence, 2026)
• The AI meeting assistants market alone was valued at $3 billion in 2025 and is projected to reach $6.28 billion by 2035, driven by over 45% of enterprises seeking sentiment analysis and decision-tracking from meeting data.
• Gartner predicts conversational AI will automate approximately 70% of customer support interactions in enterprises by end of 2027, up from 50% in 2025, making support-focused patterns among the fastest-maturing vendor categories.

RAG Assistant: Knowledge base AI and enterprise search

Product category: Enterprise AI search, internal knowledge assistant, company copilot, RAG platform

Maturity: High

Typical buyer: IT, HR, Support Operations, CIO office

The RAG Assistant category is one of the most crowded in business AI. It spans three distinct market segments.

Enterprise platform plays: Major technology companies have embedded RAG-style capabilities into their productivity suites. Microsoft's Copilot across Office 365 and SharePoint, Google's Workspace AI, and similar integrated products serve teams already within those ecosystems. Integration effort is low; vendor dependency is high.

Dedicated enterprise search with AI: Products like Glean build knowledge retrieval as a standalone product that indexes across multiple systems (CRM, email, Slack, Google Drive, Confluence) and generates answers. The value proposition is cross-system retrieval, not just within one platform.

Point solutions for specific contexts: Notion's Q&A feature, Confluence AI, Zendesk's AI knowledge base, and similar products serve RAG within a specific tool's scope. These are lower complexity but narrower coverage.

When evaluating, ask: what systems does the product index, and can it connect to all the places where your organizational knowledge actually lives?

Scoring + Routing: Predictive CRM and intelligent triage

Product category: AI lead scoring, predictive sales analytics, revenue intelligence, ticket routing AI

Maturity: High for sales lead scoring; moderate for support routing, HR screening, and other applications

Typical buyer: RevOps, Sales Operations, Support Operations

Sales lead scoring has a mature vendor category embedded in the major CRM platforms. HubSpot's predictive lead scoring, Salesforce Einstein, and dedicated RevOps analytics tools (MadKudu, 6sense in ABM contexts) all serve the Scoring and Routing pattern. The market is mature enough that default configurations produce useful results for standard B2B sales motions.

Support ticket routing AI is embedded in most major help desk platforms (Zendesk, Freshdesk, Intercom) as a native feature. The routing logic is usually simpler than sales scoring, but the category is similarly mature.

Recruiting AI scoring (resume screening, candidate ranking) is a distinct sub-market with dedicated vendors. This sub-market faces additional regulatory scrutiny around algorithmic bias in employment decisions, which affects both vendor compliance requirements and internal governance.

When evaluating, ask: does the vendor's default scoring model reflect your industry and deal motion, or will it require significant retraining before it's useful in your context?

Vision Extract: Intelligent document processing

Product category: Intelligent document processing (IDP), AP automation, OCR+AI platforms, KYC document verification

Maturity: High for standard document types (invoices, receipts, IDs); moderate for specialized formats

Typical buyer: Finance, AP/AR teams, Operations, Compliance

This category splits cleanly by document type. Standard financial documents (invoices, purchase orders, receipts) have a mature vendor market with high accuracy rates. Vendors include Klippa, Mindee, Kofax, and ABBYY, alongside AP automation platforms (Tipalti, Bill.com) that embed extraction as part of a broader workflow.

Identity document verification (passports, driver's licenses, national IDs) is a distinct sub-market used primarily by fintechs, banks, and businesses with KYC requirements. Vendors here (Veriff, Jumio, Onfido) are specialized and operate under significant regulatory frameworks.

Specialized document types specific to your industry (manufacturing inspection forms, healthcare intake forms, proprietary contracts) typically require custom training data on top of a vendor base model. No vendor has a production-ready model for your specific document format unless your format is industry-standard.

When evaluating, ask: has the vendor's model been trained on documents that look like yours, and can they demonstrate accuracy on your specific document types before you sign?

Meeting Intelligence: Conversation intelligence

Product category: Conversation intelligence, revenue intelligence, call recording + AI, sales coaching AI

Maturity: Very high

Typical buyer: Sales leadership, RevOps, Customer Success leadership

This is one of the most mature pattern categories in business AI. Gong, Clari Copilot, Chorus (now ZoomInfo), Fireflies, and Otter for Business all serve the Meeting Intelligence pattern with production deployments at scale. The core pipeline (recording, transcription, topic extraction, CRM push) is commodity. Differentiation is in coaching analytics, deal risk detection, and depth of CRM integration.

The category is also under pressure from the collaboration platforms themselves. Zoom, Microsoft Teams, and Google Meet all offer native AI meeting summaries with direct calendar and CRM integration. For teams that want basic transcription and summary, the platform-native options are increasingly competitive with dedicated conversation intelligence tools.

The decision between dedicated vendors and platform-native capabilities usually comes down to coaching depth and cross-meeting analytics. Platform-native tools summarize individual meetings. Dedicated conversation intelligence tools analyze patterns across hundreds of calls, track coaching metrics over time, and integrate deeply with CRM deal context.

When evaluating, ask: do you need meeting-by-meeting summaries, or do you need pattern analysis across your full call library?

Anomaly Agent: Multiple sub-markets with different maturity levels

Product category: Fraud detection, AIOps monitoring, security threat detection, expense anomaly detection

Maturity: High for fraud and infrastructure; moderate for business process anomalies

Typical buyer: Finance/Risk (fraud), Engineering/DevOps (infrastructure), Security (threats), Finance/HR (expense and process anomalies)

This pattern has four distinct sub-markets that rarely overlap in vendor landscape.

Fraud detection: One of the most mature AI applications in existence. Stripe Radar, Sift, Forter, and embedded fraud scoring in payment processors have been in production at scale for years. These vendors have data advantages (trained on industry-wide transaction patterns) that internal builds can't match.

Infrastructure and application monitoring (AIOps): Datadog, New Relic, Dynatrace, and Splunk all provide anomaly detection on metrics, logs, and traces. The category is mature and embedded in DevOps toolchains.

Security threat detection: SIEM platforms (CrowdStrike, Sentinel, Splunk SIEM) have anomaly detection as a core feature. This sub-market is specialized and typically owned by Security rather than IT operations.

Business process anomaly detection: Detecting unusual expense patterns, HR policy deviations, supply chain anomalies, or operational process deviations is the least mature sub-market. Some expense management platforms (Ramp, Brex) are building this in. But for non-financial business process anomalies, you're often in build territory or working with general-purpose monitoring tools adapted to your use case.

When evaluating, ask: which sub-market does this vendor actually serve, and does their training data and baseline model reflect your specific process?

Generative Research: AI research assistant

Product category: AI research assistant, competitive intelligence AI, account research automation

Maturity: Emerging. Significant variation in source handling and output quality.

Typical buyer: Strategy teams, sales (account research), marketing (competitive intelligence), analyst functions

The generative research category is young and fragmented. General-purpose AI tools (Perplexity, You.com Pro, ChatGPT with Browse) serve this pattern for public-source research. Dedicated competitive intelligence tools are proliferating but still maturing.

The key differentiation to evaluate in this category is source access. Different vendors have access to different source types: public web, news archives, financial databases, proprietary industry data, and internal document repositories. The research quality is a function of what sources the product can actually access, not just what the generation layer produces.

A second differentiation is citation fidelity. Some tools produce well-cited research with traceable sources. Others hallucinate citations or paraphrase so aggressively that the original source is unrecoverable. This matters significantly for any research that will be used externally or distributed to decision-makers.

When evaluating, ask: what are the actual sources this product pulls from, and can it demonstrate citation accuracy on a research task from your domain?

Document Review: Contract AI and beyond

Product category: Contract AI, legal AI, compliance document review, CLM (contract lifecycle management) with AI

Maturity: High for contract review; emerging for specialized domains

Typical buyer: Legal, Procurement, Compliance

Contract review AI is a mature category. Spellbook, Harvey, Ironclad AI, and LexCheck are purpose-built for legal document analysis. Larger CLM platforms (Ironclad, Conga, Icertis) have embedded AI review as part of broader contract workflow tools. The category has proven production deployments at large and mid-market companies.

The category gets thinner as you move away from standard legal contracts. Tax filing review, insurance policy comparison, regulatory compliance review in non-legal contexts, and technical document review (reviewing code for security compliance, reviewing manufacturing specs for regulatory conformance) are served by a mix of specialized tools and custom-built solutions. Few vendors have productized these applications to the same maturity as legal contract review.

When evaluating, ask: has this vendor processed documents in your specific domain, and can they show accuracy benchmarks on document types similar to yours?

Workflow Copilot: Highly fragmented by context

Product category: AI copilot, role-specific AI assistant, horizontal productivity AI, domain copilot

Maturity: High for horizontal work (writing, coding); moderate for domain-specific

Typical buyer: Varies by context: IT/Engineering for coding copilots, Operations for domain copilots, specific function heads

This is the most fragmented category in the vendor landscape. The pattern is extremely versatile, which means it's been productized in dozens of specific contexts.

Horizontal copilots: Microsoft 365 Copilot (email, documents, meetings), GitHub Copilot (code), and similar platform-level offerings serve broad horizontal knowledge work. These are mature, high-adoption products with large-scale production deployments.

Domain-specific copilots: Sales copilots (integrated in Salesforce, HubSpot, or as dedicated add-ons), support copilots (in Zendesk, Intercom), finance copilots, and marketing copilots serve specific workflow contexts. Rework's Sales AI falls into this category. Quality and integration depth vary significantly. For how these vendors stack up specifically in the AI sales ops context, the AI Sales Ops vendor landscape in 2026 maps the full competitive set.

Infrastructure for building copilots: For teams building their own domain copilot, LLM provider APIs (Anthropic, OpenAI, Google), orchestration frameworks, and vector database vendors provide the building blocks.

When evaluating, ask: how deep is the integration with the specific tool where this copilot lives? A copilot that's bolted onto a CRM it can't read is less useful than one that's native to the workflow.

Personalization Engine: Mature for e-commerce, growing for B2B

Product category: Recommendation engine, dynamic content platform, AI personalization, CDP with AI activation

Maturity: High for e-commerce; moderate for B2B SaaS and content platforms

Typical buyer: Marketing, Product, E-commerce

Dynamic Yield, Bloomreach, and Monetate serve the e-commerce personalization market with mature, high-scale platforms. The category has significant depth in product recommendations, dynamic pricing, and page-level content personalization.

In B2B SaaS, Mutiny and Intellimize focus on website personalization (different content for visitors from different companies or industries). Segment and similar CDPs (Customer Data Platforms) provide the behavioral data layer that personalization engines consume. The B2B segment is less mature than e-commerce but growing.

In-product personalization for SaaS (adapting the product experience based on user behavior) is primarily custom-built or done through product analytics platforms that enable targeted feature flags and in-app messaging (Amplitude, Mixpanel with experimentation features).

When evaluating, ask: does this vendor's personalization model handle your user structure (individual users, account-level users, anonymous visitors) and your content volume?

Autonomous Agent: Early-stage, high activity

Product category: AI agent platform, agentic workflow automation, agent framework

Maturity: Emerging. High market activity, limited proven production deployments at scale.

Typical buyer: CTO office, AI/ML engineering teams, Operations

The Autonomous Agent pattern category is the most active in terms of new vendor announcements and the least mature in terms of proven enterprise deployments. LangChain, CrewAI, and AutoGen provide frameworks for building agents. Vertical-specific agent platforms are proliferating in sales development, customer support, and software engineering contexts.

The category's maturity gap is significant: the frameworks exist, but governance tooling (approvals, audit trails, escalation paths) is still being built out. Most enterprise deployments of autonomous agents are in controlled, bounded contexts (a research agent that never writes to external systems, a code agent that operates only in a sandbox) rather than in fully agentic production workflows.

When evaluating, ask: what does the vendor's error handling and escalation infrastructure look like? An autonomous agent without a clear escalation path for cases it can't resolve is an audit risk, not a productivity tool.

"Vendor capability in AI pattern categories shifts substantially in 18 months. The meeting intelligence vendors that dominated in 2023 face direct competition from Zoom, Teams, and Google Meet native summaries in 2026. Buyers should evaluate against the pattern's capability requirements, not just current vendor positioning." (Rework Vendor Landscape Analysis, 2026)

The Pattern Vendor Map

The Pattern Vendor Map is an evaluation framework that classifies AI vendors by which of the 10 ACE patterns they serve rather than by marketing category. The framework has four evaluation dimensions: (1) category maturity (high, moderate, or emerging), (2) typical internal buyer persona (who should own this decision), (3) the single most important evaluation question for that pattern category, and (4) whether a platform-native option exists that reduces integration overhead. Using the Pattern Vendor Map before any vendor conversation cuts evaluation time by eliminating vendors whose pattern does not match the requirement and quickly identifying whether a platform-native option is viable for your context.

Rework Analysis: The enterprise AI market's $114 billion scale in 2026 has created significant vendor fragmentation, with hundreds of vendors each claiming "AI-powered" capabilities that may map to very different underlying patterns. In Rework's procurement experience, buyers who organize vendor evaluation by pattern rather than by feature list reduce vendor evaluation cycles from an average of 16 weeks to 8 weeks, because the pattern framework immediately filters out vendors solving a different problem than the one the buyer has.

Horizontal platforms spanning multiple patterns

Some vendors serve 3-4 patterns through a platform approach. Salesforce serves Scoring + Routing, Workflow Copilot, Meeting Intelligence, and Generative Research within its AI layer. HubSpot serves similar patterns within its CRM ecosystem. Microsoft 365 Copilot spans Workflow Copilot, RAG Assistant, and Meeting Intelligence.

Platform consolidation has real advantages: single contract, single data model, integrated authentication, and no cross-system data movement. It also has real risks: you accept the platform's version of each pattern, which is rarely best-in-class on every dimension.

The decision to consolidate on a platform versus buying best-in-class point solutions comes down to integration cost versus capability compromise. If your existing platform's versions of the patterns are 80% as capable as the best point solutions, consolidation is usually worth it. If you need 95% capability on a critical pattern, the best-in-class point solution is worth the integration overhead.

Applying the vendor landscape

The pattern framework makes vendor evaluation faster and more honest. Before any vendor conversation, answer:

1. What pattern is this vendor serving?
2. Is this a mature or emerging category for that pattern?
3. Is the vendor's version of the pattern suited to your data type and use case?
4. What customization will you need to do on top of the vendor's base product?

These four questions generate the right evaluation criteria for any AI vendor conversation. See Choosing the Right AI Pattern for Your Problem for pattern selection upstream of vendor evaluation. See Buy vs. Build Decision for Each AI Pattern for when the vendor landscape doesn't have what you need.

Governance requirements that affect which vendors are viable in your context are in Governance Requirements by AI Pattern. And for how vendor choices affect your multi-year roadmap, see Sequencing AI Patterns in a Multi-Year Roadmap.

The market is moving fast enough that any specific vendor assessment is stale within 12 months.

Gartner predicts that by year-end 2027, [conversational AI will automate approximately 70% of customer support interactions](https://www.gartner.com/en/newsroom/press-releases/2025-06-25-gartner-predicts-over-40-percent-of-agentic-ai-projects-will-be-canceled-by-end-of-2027) in enterprises, up from 50% in 2025. Vendor capability in the maturity curves above will shift substantially in 18 months. The pattern framework isn't. Use it as your stable evaluation lens.