Enterprise & Workflow AI

Best for: Companies with Salesforce as their CRM/service system of record that want AI agents handling customer-facing or rep-facing work — sales prospecting, case deflection, field service triage — grounded in Salesforce records and Data Cloud.

• Pricing — consumption (preferred): Flex Credits at $500 per 100,000 credits; a standard agent action consumes ~20 credits, so each action is ~$0.10. Or the Conversations model at $2 per conversation (24-hour interaction window). The two cannot coexist in one org
• Pricing — per-seat add-ons: $125/user/month for Agentforce for Sales / Service / Field Service, including unlimited employee-facing agent usage inside Salesforce
• Workflows: Service case resolution, sales rep coaching, lead qualification, field dispatching, two-way messaging across email/SMS/WhatsApp, voice agents
• Spring '26 components (GA): Agentforce Builder, Agent Script, Agentforce Voice, and Intelligent Context
• Grounding: Native access to Salesforce records, Data Cloud unification, and the Atlas Reasoning Engine; Einstein Trust Layer provides dynamic grounding, zero data retention, and toxicity detection
• Open extensibility: Spring '26 added ChatGPT integration and stronger Slack hooks; Prompt Builder + Copilot Builder let admins extend agents without code

Limitations: Strongest fit when call center / service org volume is high enough to make case-deflection meter math compelling. Be careful with Conversations pricing if your reps run dozens of short interactions per day — Flex Credits typically wins when conversations are short-and-many; Conversations wins when they're long-and-multi-action.

Best for: Organizations on Google Workspace that want AI baked into Gmail, Docs, Sheets, Slides, and Meet without a separate per-seat AI add-on.

• Pricing: Google discontinued the standalone Gemini Business / Gemini Enterprise add-ons in late 2025 and bundled the AI features directly into Workspace plans. Business Starter $8.40/user/mo, Business Standard $16.80, Business Plus $26.40 (annual). Most plans absorbed roughly a 17% price increase to cover the bundle
• Workflows: Help me write in Gmail/Docs, formula generation in Sheets, slide generation in Slides, real-time meeting notes and translation in Meet
• Models: Gemini 2.x family across the Workspace surfaces; NotebookLM for source-grounded research is available alongside
• Custom AI: Gems (saved custom assistants), AI Studio for experimentation, Vertex AI for production builds
• Grounding: Reads from your Drive, Gmail, and Calendar; results are workspace-scoped, not training data
• Privacy: Enterprise data stays under Workspace contract terms; no training on customer prompts; data residency options on higher tiers

Limitations: The bundling change is the simplest pricing story of the cohort — you don't pay extra for AI, but you do pay more per Workspace seat than you used to. Best fit when Workspace is already mandatory; weakest fit when you're trying to layer AI onto a non-Google productivity stack.

Best for: Enterprises running ITSM, HR Service Delivery, Customer Service Management, or Field Service on ServiceNow who want AI handling agent assist, summarization, and case automation inside Now Platform workflows.

• April 2026 retiering: AI now embedded across Foundation / Advanced / Prime tiers — no separate Now Assist SKU. Specific capabilities and token pools vary by product and tier
• Workflows: Incident summarization, change risk prediction, knowledge article drafting, virtual agent deflection, email triage, code generation in App Engine
• Models: ServiceNow's own Now LLMs (domain-tuned) plus partner models (OpenAI, Microsoft, Google) routed via the Now Platform; BYO models on Prime tier
• Grounding: Native access to CMDB, knowledge bases, and process records via Workflow Data Fabric and Context Engine
• Now Assist Skill Kit for custom skills; AI Agent Studio for custom agents
• Privacy: Tenant-scoped, not used for cross-customer training; standard ServiceNow compliance posture (FedRAMP, SOC 2, HIPAA paths)

Limitations: Strong on case-deflection and time-to-resolution math when ticket volume is high. Watch for token-pool overage on Advanced / Prime — AI is 'included' in the new tiers but consumption above the pool incurs overage fees on top of an already premium per-seat license. Pricing remains heavily NDA-bound; list prices for Foundation / Advanced / Prime are not disclosed publicly.

Best for: Teams that live in Slack and want fast catch-up on channels and threads, plus search that reasons over conversation history — without paying a separate AI add-on.

• Pricing: AI features are now bundled into Slack plans. Pro at $8.75/user/month gets summaries and search enhancements; Business+ at $18/user/month adds advanced AI plus SAML SSO and SCIM. The standalone $10/user AI add-on was discontinued; the previous $20/user advanced AI add-on was rolled into Business+
• Workflows: Channel and thread summaries, daily catch-up digests, AI search that answers questions from message history, recap of missed conversations
• Models: Salesforce-managed model routing under the Einstein Trust Layer; no training on customer messages
• Salesforce ecosystem: Tightening integration with Agentforce — Slack increasingly serves as the conversational surface for Salesforce agents
• Custom agents: Slack apps and the new Agentforce-in-Slack pattern let teams expose CRM agents in DMs and channels
• Privacy: Workspace-isolated, enterprise key management on Enterprise Grid

Limitations: Most measurable on huge, noisy workspaces — the catch-up summaries genuinely cut hours of scroll time per week. Less compelling on small, focused teams where everyone reads everything anyway. Easy to expense; hard to abuse pricing-wise because it's per-seat-bundled now.

Best for: Teams that already use Notion as their wiki, project tracker, and doc system — especially scrappier orgs that don't want to commit to Microsoft or Google's full stack.

• Pricing: Notion bundled AI into the paid plans in 2025. Business at $20/user/month annual ($24 monthly) includes unlimited Notion AI; Enterprise is custom-quoted with advanced security and compliance. Free and Plus tiers no longer have an AI add-on path for new users
• Workflows: Document drafting, summarization, action-item extraction, database autofill, content translation, meeting-note rewrites
• AI Agents (GA Sept 2025): Persistent agents that act on your Notion content; Custom Agents arrived with Notion 3.3 in February 2026
• Ask Notion: Cross-workspace search and Q&A that reaches into connected sources — Google Drive, Slack, GitHub — not just Notion pages
• Models: Mix of Anthropic Claude and OpenAI GPT models routed by Notion; no training on customer content
• Privacy: Workspace-scoped, with admin controls over which AI features are enabled and which connectors are allowed

Limitations: Best when Notion is the primary knowledge surface — if your docs are scattered across SharePoint, Google Drive, and Confluence, Notion AI's grounding story is much weaker. Strong fit for product, design, and engineering teams; weaker fit for finance, legal, and ops orgs anchored in Microsoft or Google.