Jan 2026 Release Notes

Jan 2026 Release Summary

💡

The January 2026 release is a foundational intelligence and collaboration upgrade for Sapience. This release introduces persistent Agent Memories, voice-based output, richer project visualization, and significant file system performance improvements, while also strengthening collaboration through shared conversations and group chats.

In parallel, this release hardens the Gen 3 Agent Engine, improves system reliability, and delivers meaningful usability enhancements across Notes, Projects, Files, and Agents.

Key Changes

New Agents: New dedicated Data Analyst Agent, plus better Coding Agent availability on Gen3 Agents

Projects Upgrade: Projects Improvements (Eisenhower Matrix, Kanban Board, Task IDs)

Scheduled Work: Set up Sapience Agents to work on a schedule, in the background, 168 hours a week.

Documentation Portal: 200+ pages of good documentation for every part of the system, including frequently asked questions, security and integration guides and more.

Docs AI Agent: try the AI Agent on top of Docs. Its great!

Memory: Persistent Agent Memories across Gen 3 agents

Group Chats: Conversation Sharing & Group Chats

Vision: Image Recognition / OCR Analysis. Generate Images out.

Localization: Product interface localised in Arabic, German, French, Spanish and Hindi.

Agents: New Agent Browser Grid View & Features Panel

File improvements: Major File System Enhancements (Floating Chat, Reprocessing, Zstd compression)

Note templates: Global Note Templates (Meeting Notes, OKRs)

Voice Output Mode for hands-free consumption of AI responses Experimental

Gen 3 engine improvements: reliability, system refinements, and bug fixes

*New Eisenhower Matrix view of Tasks inside of Projects.*

*New Kanban view for managing Projects and Tasks.*

*New Documentation portal + AI Help Agent that is trained on the 50+ articles.*

1. Projects Changes

Projects now include two powerful visual planning tools for prioritization and execution.

1.1 Eisenhower Matrix

Key capabilities

Each Project can now be viewed as an Eisenhower Matrix (Urgent vs. Important), which can also be dragged and dropped interactively:

Q1 – Urgent & Important (Do Now)

Q2 – Important, Not Urgent (Schedule)

Q3 – Urgent, Not Important (Delegate)

Q4 – Neither (Eliminate/Minimize)

Tasks and goals already defined in Projects & Tasks can be dragged between quadrants; urgency/importance flags are updated under the hood.

Why it matters

Gives executives a strategic prioritization lens directly inside Projects, building on the same work‑graph that already supports Related Resources and fullscreen grids.

1.2 Kanban Board Views

Key capabilities

Drag-and-drop task statuses

To Do / In Progress / Blocked / Completed

Cards show titles, assignees, due dates, and related Notes/Files; board views reuse Sapience’s existing grid/fullscreen infrastructure.

Optional WIP limits per column encourage focus and prevent overload, aligned with Kanban best practices drawn from industry usage

Why it matters

Couples the Matrix’s “do the right things” perspective with a Kanban view for “do things right”, turning Projects into a complete prioritization + execution surface.

1.3 Additional Project Enhancements

Task IDs

Dynamic numerical IDs per project

Auto-renumbered when tasks are deleted

Resizable Columns

Files, Notes, Projects, and Tasks lists

Floating Project Chat

Movable project agent chat window

Persists across project tabs

Why it matters

Large projects become easier to reference, analyze, and manage visually.

2. Agent Platform and Intelligence

2.1 New Data Analyst Agent & Coding Agent Tool

There is now a Gen3 global agent, the Data Analyst Agent. It will fire up a sandboxed container and use the Sapience Coding agent behind the scenes to analyze data sets (e.g. Excel or JSON data). This is great for analyzing any tabular data files, and is optimized for .xls, .xlsx, .csv and .json.

Note also that any Agent that advertises the Coding Agent Feature on its Feature Card can dynamically hane tabular data files when attached to a conversation. To try these out either:

Attach an excel file to any Gen 3 Agent that shows the Coding feature; or

Do it in the Data Analyst Agent

It will write code to do numerical analysis, and is great at generating static images as well as interactive dashboards. If you ask it to create a dynamic dashboard as a standalone HTML file, and tell it what you want the dashboard to depict (the more detail the better) then you can create 3D models, histograms, data-dense heatmaps and pretty much any kind of static or interative visual that you can describe (with reference to the data you upload).

2.2 Agent Memories - Cross-Agent Memory Layer

Agent Memories introduce system-level, persistent memory across all Gen 3 agents. Agents can now store, retrieve, and reason over shared knowledge such as facts, preferences, events, and instructions—across conversations and across different agents.

Key Capabilities

Cross-agent memory recall (shared knowledge context)

Persistent Knowledge Graph for user facts and preferences

Natural-language memory triggers

“Remember that …” to store

“What do you know about me?”/ “List your memories” to retrieve.

Four-tier memory architecture:

Session memory (temporary)

Chat-level context

Persistent agent memory

System-wide shared memory

Contextual lookup to refine responses automatically

Why it matters

Agent Memories turns Sapience into a compound memory system rather than a set of stateless chats. This enables continuity (e.g., time zone, role, client preferences) and cross-agent collaboration without repeated prompting, which is particularly valuable for executives and teams that frequently switch between different specialist agents.

2.3 Voice Output Mode (Experimental)

Voice Output Mode enables agents to read responses aloud, allowing users to listen to summaries, reports, and answers while multitasking or working hands-free.

Key Capabilities

One-click “Listen” on agent responses via a speaker icon or “Listen” button to start text-to-speech playback for answers, summaries, and reports.

Playback controls: play, pause, resume, and seek within the audio, following standard voice UX patterns.

Configurable playback speed with user presets (e.g., 0.75× to 2×).

Voice selection & language support from Settings, allowing personalized voices and accents within admin policy.

In-line text highlighting synchronized with audio playback for better comprehension and accessibility.

Hands-free friendly design with keyboard-accessible controls and optimization for non-mouse usage.

Admin controls for organization defaults (whether Voice Output is enabled, which voices and languages are allowed).

Why it matters

Voice Output Mode closes the loop between Sapience’s existing dictation inputs and consumption of outputs. Executives can now listen to complex summaries, file analyses, and weekly briefings while commuting or context-switching, reducing screen load and supporting accessibility needs in high-intensity workdays.

2.4 Agent Browser – Grid View & Features Panel

The Agent Browser has been redesigned to improve discoverability and transparency of agent capabilities.

Key Capabilities

Grid-based agent browser (replacing list view)

Unified sidebar navigation with View All

Search and filtering across agents

Per-agent Features Panel showing:

Enabled tools

Skills (search, memory, coding, uploads, etc.)

Scope (Global / Org / User)

Special configurations

Dynamically generated capability panels (not static checklists)

Why it matters

Users can immediately understand what each agent can do, enabling better agent selection and fewer workflow surprises.

3. Scheduled Jobs & Background Work

Major upgrade to Sapience and its ability to work 168 hours a week as your digital co-worker. See this article for details:

Scheduled Jobs & Agents

4. Collaboration and Conversations

4.1 Conversation Sharing & Group Chats

Conversations can now be shared with other users or via public/external links, transforming isolated chats into collaborative, multi-user threads.

Key Capabilities

Share specific conversations with other Sapience users or create a public/external link to a thread.

Supports multi-user conversations with clear participant identification in the thread.

Adds visual indicators for shared conversations in the sidebar (share icon + highlight).

Updates the ChatHeader to show:

Shared status.

Refresh button.

Conversation owner and participants.

Lays groundwork for multi-agent, multi-user collaboration in a single shared context.

Why it matters

Previously, many Sapience chats were effectively single-user sandboxes. With conversation sharing and group chat, teams can now co-develop analyses, strategies, and documents in a shared AI thread, making it much easier to move from individual insights to collaborative decisions.

5. Files and Content Intelligence

5.1 File System Enhancements

Major upgrades improve multitasking, recovery, and performance when working with files.

Key Capabilities

Floating File Chat — movable chat windows

Reprocess / Re-transcribe File action to regenerate transcripts or raw text

Zstandard (Zstd) compression for files ≥25MB

3–5× faster document uploads

20–30% faster media uploads

Concurrent chats (file chat + project chat + group chat)

Why it matters

Users can work with multiple files simultaneously, recover from rare processing issues, and upload large files significantly faster.

5.2 AI Vision - Image Recognition and OCR Analysis

Sapience agents can now extract and reason over text from images using OCR, making images first-class inputs.

Key Capabilities

New Image OCR pipeline for image files (PNG, JPG) and image‑based PDFs:

Automatically extracts embedded text on upload and stores it alongside file metadata.

Enables search and chat over screenshots, scanned contracts, whiteboard photos, and receipts, via File Info Card and Single File Chat.

Designed with multi‑language extensibility in mind, complementing Sapience’s localization direction.

Why it matters

Elevates previously opaque image assets into searchable, analyzable content, tightly integrated with Sapience Agent and Projects.

Reduces the need for external OCR tools and manual re‑typing of image‑based information.

Take a photo of a diagram or whiteboard and put it into chat with Gen3 Agents. Its an amazing productivity boon.

5.3 Images Out - Generate Image tool

Ask Sapience to generate an image, a diagram, flowchart, or a picture of a purple unicorn.

It can do it now! This works as a terrific complement to both Visual Mode, and the AI Vision feature.

6. Notes & Knowledge Management

6.1 Global Note Templates

Sapience now ships with Global Note Templates – standardized Meeting Notes and OKR templates available to every user in every organization, prominently surfaced in the note creation flow.

Included Templates

Meeting Notes

OKR (Objectives & Key Results)

Voice Note

Project Overview

Key Capabilities

Globally available templates across all organizations and accounts.

Built‑in templates for Meeting Notes and OKR Planning with structured sections for agenda, attendees, objectives, and key results.

Previews in the note creation modal and prominent placement in the + New Note dropdown.

Encourages creation and management of custom templates on top of the global set.

Why it matters

Accelerates creation of structured notes for common workflows (meetings, OKR planning).

Demonstrates the value of reusable templates and encourages teams to adopt their own patterns.

7. Other Enhancements

7.1 New Documentation Portal

New portal (you’re looking at it now!) available at https://help.sapiencecloud.ai and linked from the main website www.SapienceCloud.ai

Also available inside the app under “Docs & Resources” > “Documentation”

50 articles, with an AI Agent over the top that you can chat with.

Great AI agent over the top of the docs which makes it easy to find what you need, but grounded in all the Documentation

7.2 Improved Conversations List, Pinning, and Search

Key capabilities

Refined Conversations list:

Clear titles, last‑message previews, timestamps, and unread indicators.

Layout aligned with other Sapience object lists for familiarity.

Pinned conversations section at the top of the list for high‑priority threads (e.g., board‑prep, key project agents).

Integrated search and filters across conversation titles and content, consistent with how Alt+F Global Find surfaces chats and messages.

Improved continuity: reopening a conversation restores you to your last read position, making it easier to resume long‑running workstreams.

Why it matters

Reduces friction in navigating a growing number of chats and makes Sapience’s conversational layer feel more like a first‑class work hub than a simple history log.

7.3 Gen 3 Agent Engine Enhancements

LLM System Instruction Logging

Optional capture of final instructions sent to the LLM

Supports debugging and troubleshooting

Message Bus Error Handling

Centralized exception handling

Integrated with Sentry, EmailNotifier, and Events backend

Why it matters

Improves platform reliability, observability, and production-grade stability.

8. Other

Default Web Research Agent restored for web searches

Project Wizard simplified (file upload step hidden)

“Duplicate” action deprecated

Save as template enabled for all user roles

Sapience Agent now confirms before invoking heavy tools

Resumable uploads enabled by default (≥25MB)