Overview
Start with What Hana Is (And How To Use Her) for a plain-language walkthrough.
If setup or output quality feels off, contact us early instead of waiting. We can help with onboarding, prompt quality, memory setup, connectors, and workflow design.
Start here: Support Channels, FAQs, and Troubleshooting.
What Hana Is
Hana is an information agent for Google Chat and Google Workspace-centered collaboration.
Unlike a generic chatbot, Hana continuously maps team context to action: she ingests organizational knowledge, retrieves what matters, and helps teams execute with shared context.
Hana helps teams:
- Understand context from ongoing conversations.
- Retrieve and summarize relevant organizational knowledge.
- Execute common operational workflows (meetings, reminders, tasks, reporting, memory retrieval).
- Turn chat intent into action with tool-aware responses.
- Reduce context switching by bringing knowledge + execution into one chat workflow.
Where Hana Works Best
Hana is most effective when your team uses:
- Structured threads in Google Chat.
- Clear prompts with time range and scope.
- Memory sources (docs, policies, SOPs, links, workspace knowledge).
Clear Use-Case Map
Think of Hana as a 3-step information agent:
- Collect context: from thread, space, memory, docs, and connected sources.
- Reason over context: summarize, compare, identify risk, find gaps.
- Execute next step: create tasks/reminders/polls/meetings and publish updates.
If a user is unsure where to start, this default prompt works in almost every team:
@Hana summarize this thread, list decisions, and create action items with owner and due date.
How Hana Works (Visual)
Core Capabilities
- Chat-native information-agent assistance in Google Chat.
- Tool routing for workflows like calendars, reminders, tasks, docs, reporting, and memory operations.
- OpenAI-powered image generation and editing workflows in chat (create, recolor, replace, outpaint, mask-based edits).
- Memory-augmented answers grounded in your team context.
- Web-informed responses for queries that need external context.
- Multi-language support for prompts and responses.
Context Model (How Responses Stay Relevant)
Hana combines multiple signals when answering:
- Current message intent.
- Relevant recent chat context.
- Thread context when invoked inside a thread.
- Enabled memories and connected knowledge sources.
- Time expressions in your prompt (for example, "today", "this week", "from Monday to Wednesday").
For a deeper breakdown, see History Scoping.
Prompting Patterns That Work
Use explicit prompts with these three ingredients:
- Objective: what outcome you need.
- Scope: where to look (
this thread,this space,emails,memory). - Time window: exact range when relevant.
Examples:
@Hana summarize decisions in this thread and list action items with owners.
@Hana check calendar availability for @Jane and @Akash this Friday and suggest 3 slots.
@Hana review the attached doc and list missing sections for launch readiness.
@Hana report daily_standup updates from today 9am-5pm by owner.
Recommended First 30 Minutes
- Add Hana to your working chat space.
- Run 2-3 real prompts from active team work.
- Add at least one memory source (policy doc, SOP, or project context).
- Move ongoing discussions to threads for cleaner continuity.
- Save recurring prompt templates in your team playbook.
For a detailed day-by-day rollout, use Quick Start: First Week Plan (Teams).