Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.teeem-ai.com/llms.txt

Use this file to discover all available pages before exploring further.

​
Product

Teeem AI is an enterprise team AI agent that runs natively in Slack and web chat. It pulls together data scattered across your ERP, groupware, internal docs, and email into a single shared context for an employee’s question, then carries through to reports, insights, and approvable actions — all within the user’s permissions. Unlike general-purpose tools like ChatGPT or Copilot that stop at a text answer, Teeem AI grounds answers in your company data and goes through to executable work.
ChatGPT-style generative AI replies with text and stops there. Teeem AI is different.
ChatGPT etc.Teeem AI
Output shapeTextText + tool execution + file generation + system calls
Data connectionLimited to training dataCompany ERP, mail, calendar, apps, e-commerce
PermissionsNot bound to user identityPer-user RBAC + ABAC + channel ACL
AuditConversation history (manual)Tamper-evident hash chain (5 years)
CitationsGenerally noneEvery grounded answer comes with sources
DeploymentSaaS onlySaaS + on-prem (negotiated)
In one line: an AI that goes from answering to executing.
Yes. Teeem AI learns your company’s report templates, approval flows, and work patterns. The LLM itself is not trained on your data (privacy and confidentiality). Instead, company context is defined explicitly in two places:
  • Knowledge Base (KB) — upload policies, templates, FAQs as PDF / Excel / HWP
  • App Pack — model your data with app.yaml (models, fields, relations)
Once defined, this company context is automatically applied to every response. Starting with 10–20 core documents delivers immediate value.
Yes — from 10-person startups to large enterprises, regardless of size.For smaller teams:
  • No dedicated IT team required
  • Standard SaaS rollout in 30 minutes
  • Phased adoption — turn on only the channels and connectors you need
  • Token-usage based pricing — pay only for what you use
For larger enterprises, SSO, on-prem deployment, and custom integrations are layered on top. Talk to sales and we’ll design a fit for your scale.
No dedicated IT team is required. Standard SaaS rollout needs only:
  • One admin’s email
  • A Slack workspace (or willingness to use web chat)
  • (Optional) ERP / groupware API access for connectors
Initial provisioning (about 30 minutes) is handled by your sales / onboarding contact. After that, you can manage users, channels, and documents directly in the admin console.If you have an IT team, additional work is typically SSO integration, security policy customisation, and monitoring dashboard wiring.
Google Gemini is the primary LLM. Simple/medium queries route to Gemini 3.1 Flash Lite; complex work uses Gemini 3 Flash Preview. Conversation summarisation uses Groq (gpt-oss-120b) as a helper. Anthropic Claude is no longer used.
No. Teeem AI calls Google Cloud Vertex AI under enterprise terms in a configuration that does not use customer data for model training. Data is not exposed to external LLM providers for training purposes.
No. Every customer has its own PostgreSQL instance, DynamoDB tables, and file system. There is no code path that can reach another customer’s data.
Teeem AI is Korean-native by design.
  • PII masking — auto-detects 10 Korean PII types
  • Search tokenizer — BM25 with Korean particle stripping
  • Prompt-injection defence — recognises Korean patterns
  • HWP/HWPX native processing — Korean government and public-sector standard
  • Noto Sans KR optimisation throughout the UI
Yes. Beyond Korean, English and Japanese are supported. Intake forms, chat, and email notifications all localise. Set the default language per organisation; users can switch in the chat as well.

​
Security & compliance

Yes. The technical and managerial safeguards required by Korea’s Personal Information Protection Act (encryption, access control, audit logs, deletion-request handling) are all implemented. Compliance evidence packages can be exported in one step.
Yes. AI-generated content is automatically marked in text, documents, and HTTP response headers. Watermarking should be on by default for Korean-market operations.
Teeem AI itself is a certified SaaS service. Many of the certification controls (access control, encryption, auditing, PII handling) generate evidence automatically that customers can use during their own certification work. Ask sales for the detailed mapping.
The standard deployment is AWS-based SaaS. On-premises and private-cloud deployments (e.g. Korean public-sector environments) are possible for larger organisations on request. Contact sales.
Yes. Every audit entry is linked via SHA256 hash chain, so altering one entry invalidates every entry that follows. The integrity verification API confirms tamper-evidence at any point in time. Default retention is 5 years.

​
Operations

  1. An organisation domain (or use the {org}.teeem-ai.com we provision)
  2. A Slack workspace or willingness to use web chat
  3. The email of one administrator
Initial provisioning takes about 30 minutes; you can start using the product right after.
Standard infrastructure handles 1,000+ concurrent active users per organisation. Larger organisations get custom infrastructure sizing.
Token-usage based by default, with monthly flat + over-usage options also available. Token rates vary by model. For an exact quote, contact sales.
If a built-in connector exists (Douzone, Wehago, Cafe24, Naver Smartstore, Hometax, NICE, Customs), turn it on. Otherwise, supply an OpenAPI spec and tools are auto-generated. See connectors.
Coming soon. Native apps for macOS and Windows are in active development — system tray presence, global shortcuts, push notifications, and offline cache. For release date and beta access, contact sales.The desktop apps inherit the same permissions, audit trail, and PII protections as web chat. Designed so no separate security review is required.
Every service is protected by automated health checks and canary rollouts. ECS auto-rollback, wave-staged deploys, and hourly drift checks catch issues before customers notice. SLAs are negotiated per contract.

​
Capabilities

Inventory, orders, task tracking, internal wikis, KPI dashboards, sponsorship management — most spreadsheets and Notion workflows can move into an app pack. See Apps.
Teeem AI uses a confidence gate. If confidence is low, the agent asks a clarifying question instead of guessing. Answers also come with citations so users can verify.
Yes. Files larger than 8 MB are automatically processed asynchronously, so the chat doesn’t block. Results are delivered back to the chat when ready.
Browserbase stealth mode handles bot-detection-tough Korean sites (Coupang, Saramin) reliably. Always honour the target site’s terms of service and robots.txt.
Three paths.
  1. Auto-suggested — every Monday at 09:00 KST the system proposes skills from repeated work patterns
  2. Agent Builder — admins author skills directly in the console
  3. Sales request — complex workflows get scoped together with your onboarding contact
See Skills.

​
Need more help?

Sales & onboarding

official@flowos.work