On January 20, 2026, Notion shipped the first release round of 2026, version 3.2, themed around fitting Notion into your world better. The update pushes AI onto phones: AI Notes starts a one-tap recording that keeps transcribing, and the full desktop Notion Agent now runs on mobile. The more consequential change is model choice — users can switch between GPT-5.2, Claude Opus 4.5, and Gemini 3, and switching keeps the existing context intact.
For product teams, the significance is not any single feature. It is the confirmation of a trend: AI has moved from premium add-on to table stakes in productivity tools, and “which foundation model am I talking to” has shifted from an architecture decision to a user-facing setting.
Mobile AI Notes and the Full Agent
AI Notes on mobile starts transcription with one tap and keeps running if you switch apps or lock the screen; it then generates clear summaries, action items, and shareable docs in seconds. This targets the most practical pain point of mobile capture: meeting recordings that die the moment the phone leaves the foreground. Background recording is now a stated spec, not a hope. Combine that with summaries, action items, and shareable docs generated in seconds after the meeting, and the meeting-notes workflow no longer needs a separate recording app or a manual upload step.
Notion Agent arrives on phones with the full desktop capability: build forms, create databases, and search company knowledge, with agents handling to-do items in the background while you move on. An agent graduating from desktop tool to pocket assistant is the strongest signal in this release — Notion is clearly treating mobile agent capability as its 2026 battleground.
Pick Your Model, Keep Your Memory
Version 3.2 lets users choose GPT-5.2, Claude Opus 4.5, or Gemini 3 directly, or let Auto select the best model for the job. The official framing is blunt: “Models will continue to change, but your memory in Notion doesn’t.”
That sentence names the core design problem of multi-model products. Models are interchangeable commodities; user context and memory are the asset. Once switching models no longer loses context, model selection demotes itself from a one-time technical bet to a preference toggle. Convenient for users — and for the platform, it builds the competition between three model vendors directly into its product surface.
AI Usage Analytics for Enterprise
Enterprise plans gain AI usage analytics under Settings → Analytics → AI. Admins can see who uses AI most, which features actually drive engagement, and where it delivers value.
This is what AI features entering the enterprise procurement cycle look like: first the capability, then the demand for quantifiable evidence of use. For IT and operations teams that must justify AI spend to leadership, these dashboards will keep appearing — and quietly become a baseline requirement in vendor selection.
People Directory, Jira Integration, and Multi-Account Mail
People Directory is a pre-built, customizable database of everyone in the workspace, with the option to import from HR systems; open it by searching “People” or via Settings → Members → View People Directory. The Jira integration syncs dev work into Notion and lets you layer on properties — quarterly goals, customer feedback, revenue impact — into combined dashboards, while Enterprise plans can edit Jira fields directly from Notion with automatic sync-back. Notion Mail now manages multiple email addresses under one account, switching between personal and work inboxes.
Performance and MCP Updates
On performance, Windows pages open 27 percent faster, the Mac desktop app is up to 11 percent faster, and pages overall load in roughly one-third less time. Enterprise MCP received upgrades: audit logging and multi-database queries are in, with controls for external AI tool connections coming soon. Notion also recapped 2025 as a year of 90-plus shipped features, including Agents, AI notes, enterprise search, and offline mode.
For developers, MCP audit logs and external-connection controls show Notion taking the governance question of enterprise data exposure to third-party AI tools seriously. Audit logging in particular gives administrators a trail of what agents actually queried, which is the precondition for letting agent features loose on real workspaces. For teams, this release moves AI from novelty to daily operations — and the next battleground is clearly mobile, plus measurable enterprise value.
Sources
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.