Enterprise AI Reference
This content was developed with AI assistance and is regularly reviewed for accuracy.
Snapshot: October 2026. Model names, prices, and product features change monthly; What's New in AI is the maintained chronology. Vendor product names on this page change often; check the vendor's site for current naming.
This page collects the outside resources enterprise AI leaders reach for most: the research worth citing, the firms that sell advice, the standards and industry rules that shape governance, and the platforms large organizations build on. It is a lookup page, not a lesson. Each section points back to the page in this track that tells you what to do with the resource.
Use this page to:
- Find the 2026 surveys behind the numbers leaders quote, with their dates
- Choose which governance standard to align your program with
- Check which industry-specific rules apply before you deploy
- Orient yourself among enterprise AI platforms before an evaluation
Research worth citing (2026)
These are the sources behind most AI-at-work statistics in 2026. Cite the edition and date, because figures move year to year and many circulate through secondary summaries.
| Source | Edition and date | What it is good for |
|---|---|---|
| McKinsey, The State of AI | "On the road to ROI," August 2026 | Adoption (89% use AI in at least one function), flat earnings impact (37%), agent scaling by company size |
| Deloitte, State of AI in the Enterprise | January 2026, 3,200+ leaders in 24 countries | Skills as the top barrier, workflow redesign (84% have not), agent governance maturity (21% mature) |
| Microsoft Work Trend Index | May 5, 2026, 20,000 AI-using knowledge workers | Organizational versus individual drivers of impact, manager effects, agent growth in Microsoft 365 |
| Stanford AI Index | April 13, 2026 (ninth edition) | Population adoption, documented incidents, investment, public opinion |
| Gallup workplace AI indicator | Quarterly; Q2 2026 | Share of US workers using AI at work, frequently, and daily |
| Anthropic Economic Index | January, March, and June 2026 reports | Which occupations and tasks use AI, and how much is delegated versus assisted |
| Menlo Ventures, State of Generative AI in the Enterprise | December 2025 | Enterprise spend and the share of deployments that are true agents (16%) |
| Gartner agentic AI forecast | June 25, 2025, still operative | The "over 40% of agentic projects canceled by end of 2027" prediction |
Several of these could only be confirmed through secondary coverage for this snapshot, and Evaluate and Measure explains how to read the two most-quoted failure statistics, Gartner's forecast and MIT NANDA's "95%," without overstating them.
Industry analysts
Analyst firms sell vendor comparisons and market research behind paywalls; their headline predictions circulate free in press releases.
| Firm | Known for | Best used for |
|---|---|---|
| Gartner | Magic Quadrants, Hype Cycles, market guides | Shortlisting vendors; checking where a category sits on the hype curve |
| Forrester | Wave reports, playbooks | Customer-experience and business-strategy framing |
| IDC | Market sizing and spending forecasts | Budget benchmarks and market context |
| CB Insights | Startup and funding tracking, AI 100 | Spotting emerging vendors and acquisition risk |
| MIT Technology Review | Independent reporting and sponsored insight reports | Trend context from outside the vendor ecosystem |
Treat analyst predictions as hypotheses. A forecast tells you what a firm expects; your own pilot data tells you what happened.
Consulting firms
Large firms offer AI strategy, implementation, and change programs. They are most useful for enterprise-wide transformation, regulated industries, and organizations without internal capacity; for a single team's adoption, the methods in this track usually suffice.
| Firm | AI practice | Typical strength |
|---|---|---|
| McKinsey | QuantumBlack | Enterprise strategy and the annual State of AI survey |
| Deloitte | AI and data consulting | Governance, regulatory compliance, workforce transformation |
| Accenture | Data and AI services | Large-scale implementation and process redesign |
| PwC | AI and analytics | Responsible AI, risk, and the annual AI Jobs Barometer |
| BCG | BCG X | Strategy and value capture |
| IBM Consulting | AI consulting | Hybrid-cloud integration and IBM platform work |
Several of these firms are also training partners in vendor programs; Anthropic's October 2026 Frontier Academy, for example, lists Accenture, Deloitte, and McKinsey among partners for training deployed engineers. Ask any consultant how they will measure value, and hold them to the method in Evaluate and Measure.
Governance standards and frameworks
Standards give your program a recognized structure, which helps with auditors, customers, and procurement questionnaires. You rarely need more than one primary framework.
NIST AI Risk Management Framework. The US government's voluntary framework, organized into four functions: Govern, Map, Measure, Manage. Free, widely referenced, and a natural default for US organizations and federal contractors.
ISO/IEC 42001 and ISO/IEC 23894. International standards for an AI management system and for AI risk management respectively. Choose these if customers or regulators expect certifiable, ISO-style management systems.
IEEE AI standards. Technical and ethical design standards, most useful to engineering teams building AI into products.
The EU AI Act's risk tiers. Not a voluntary framework, but its four tiers (unacceptable, high, limited, minimal) are the most common lens for risk-tiering even outside the EU.
Govern: Frameworks, Policy, and Compliance shows how to size governance to your organization, and the policy template turns any of these frameworks into rules people can follow.
Regulations and industry rules
The cross-industry regulatory picture (EU AI Act phases, the Digital Omnibus delay of high-risk obligations to December 2027 and August 2028, US state laws, federal executive orders) is mapped in three tiers on Govern, with a quick-reference table in the checklist. Sector rules often matter more day to day, because regulators enforce them against AI use whether or not an AI-specific statute exists.
| Sector | Rules and guidance to know |
|---|---|
| Financial services | Federal Reserve SR 11-7 on model risk management; FFIEC guidance for US banks; European Banking Authority guidance on machine learning in credit models; fair-lending and consumer-protection law |
| Healthcare | FDA rules for AI and machine learning in software as a medical device; HIPAA business associate agreements and minimum-necessary standard; WHO guidance on AI ethics and governance in health |
| Employment | Anti-discrimination law applied to hiring and evaluation tools; EU AI Act classification of employment uses as high-risk |
| Safety-critical systems | ISO 26262 functional safety for vehicles; RTCA DO-178C for airborne software |
| Data in general | GDPR, CCPA/CPRA, and sector privacy rules on lawful basis, minimization, and automated decisions |
Vendor terms are a layer of their own. Anthropic's usage policy effective November 12, 2026, for example, requires a qualified human reviewer and notice to the affected person where AI use affects health, legal rights, finances, livelihood, or essential services.
Enterprise AI platforms
Large organizations usually build on one cloud platform and buy one or two assistant suites. The landscape below is orientation for an evaluation, not a recommendation; product names in this category change frequently.
| Provider | What enterprises use it for |
|---|---|
| Microsoft | Microsoft 365 Copilot inside Office apps; Azure AI Foundry for building with models; Agent 365 (generally available May 1, 2026) for agent identity, inventory, and governance |
| Google Cloud | Gemini models and agent-building tools on Google Cloud (the platform long known as Vertex AI), plus Gemini in Workspace |
| Amazon Web Services | Amazon Bedrock for access to multiple model families; SageMaker for custom machine learning |
| IBM | watsonx for model development, data, and governance tooling in regulated and hybrid environments |
| Salesforce | Agentforce agents for sales, service, and marketing inside the CRM |
| Anthropic and OpenAI | Claude Enterprise and ChatGPT Enterprise for organization-wide assistants with admin controls, connectors, and agent features |
For model and assistant comparisons, current plans, and tool directories, use the maintained resource pages rather than this one: AI Tools and Platforms, ChatGPT vs Claude vs Gemini, and Development and APIs. Score finalists with the weighted method in Evaluate and Measure.
Where to start, by role
Executives. Read the track overview, commission a baseline of current AI use, name an accountable owner, and ask for one measured pilot before a broad rollout.
IT and security leaders. Build the inventory and approved list in Govern, set connector and agent permissions using the checklist, and define the evaluation criteria for new tools.
Business and operations leaders. Pick the first use case and run it through Evaluate and Measure; plan the people side with Drive AI Adoption.
HR and learning teams. Design the program and the training record in Build an AI Literacy Program.
Compliance officers. Map your jurisdictions with Govern and adapt the policy template; for change models such as Kotter's eight steps, ADKAR, and McKinsey's 7S, see how ADKAR is applied in Drive AI Adoption.
Key takeaways
- Cite 2026 research by edition and date, and treat analyst forecasts as hypotheses to test against your own data
- Use consultants for enterprise-wide or regulated transformations, and hold them to a measurement method
- Pick one primary governance framework, usually NIST AI RMF or ISO/IEC 42001, and use the EU AI Act tiers as a risk lens
- Check sector rules before deployment; regulators apply existing law to AI use
- Orient among platforms here, then compare and score them with the maintained resource pages and the evaluation method
Next steps
Continue to: AI for Teams and Leaders to choose your next page in the track.
- Setting up governance? Go to Govern: Frameworks, Policy, and Compliance
- Comparing assistants for a rollout? Read ChatGPT vs Claude vs Gemini
- Building agents on these platforms? Start with AI Agents Overview