Skip to main content
Post-Investment
中文EN

Products

ZHIDOU Enterprise AI Platform

An on-premises enterprise AI platform that turns knowledge scattered across documents, systems and people into assets you can talk to and act on.

On-prem: data stays inside your network20+ document formatsSwitch between leading modelsDingTalk, WeCom and Lark bots
Book a consultationSign in(opens in a new window)
ZHIDOU knowledge base interface (Chinese)

Why AI is hard for companies

Unsure how to use it, wary of the risks, unclear on the value

Data security

Public AI services put sensitive data at risk and complicate audits.

Scattered knowledge

Knowledge sits in email, documents and systems; finding it takes time.

Disconnected systems

ERP, CRM and OA data are siloed; cross-system questions need IT.

Costly trial and error

Building AI on your own wastes time and budget.

Skills gap

Most small and mid-sized companies have no AI team.

Hard to measure

Training results and knowledge value are hard to track.

The platform

One platform for the main enterprise AI use cases

AI chat + knowledge base + skills and automation + training and exams + GEO + chat bots + scheduled tasks.

Knowledge-driven

Documents, databases and experience become structured knowledge AI can use.

AI chat

Leading model providers, multi-agent answers with sources.

Automation

Skills and scheduled tasks keep AI working in the background.

Secure

Fully on-premises, data never leaves your network, full audit log.

Knowledge base

Turn company knowledge into answers

Upload contracts, policies, manuals and reports; staff ask questions and get answers with sources.

  • 20+ formats

    PDF, Word, Excel, PowerPoint, Markdown, HTML and image OCR.

  • Hybrid search

    Semantic vectors, keywords, fusion ranking and re-ranking.

  • Knowledge graph

    Entities and relations extracted automatically; conflicts flagged.

Knowledge base interface

Skills and automation

AI that does the work, not just answers

Package repetitive work as skills that run on triggers or schedules and connect to your systems.

  • Three skill types

    API calls, prompt templates and scripts, chained in multiple steps.

  • Scheduled tasks

    Collect data and send daily, weekly and monthly reports by email, DingTalk, WeCom or Lark.

  • System integration

    MCP support; works with ERP, MES and other business systems.

Skill management interface

Training and exams

From knowledge base to trained staff

Courses and exams are generated from your own knowledge base.

  • AI-built courses

    Outlines, lessons and question banks generated automatically.

  • Smart exams

    Five question types, AI-assembled papers, anti-cheating, auto-grading.

  • Certificates and dashboards

    Certificates issued on passing; progress and scores at a glance.

Training center interface

GEO

Get your brand seen and recommended by AI search

Track how DeepSeek, Qwen, Kimi and other AI search tools mention and cite your brand, improve content and verify the results.

  • Visibility tracking

    Mentions, sentiment and citation accuracy across platforms.

  • Content optimization

    AI-friendly content generated from your knowledge base.

  • Verification

    Simulated searches track the effect over time.

GEO platform interface

Deployment and service

Runs on your own servers

Windows Server, Linux or Docker. Recommended 4 CPU cores, 16 GB RAM and 100 GB SSD; local models need a GPU.

  1. 01

    Discovery

    Your use cases and data today.

  2. 02

    Installation

    Installed and set up on your servers.

  3. 03

    Knowledge import

    Core documents loaded; search and permissions configured.

  4. 04

    Agents live

    Assistants, skills and bots configured.

  5. 05

    Ongoing service

    Updates, online support and troubleshooting.

Industries

Built for many industries

01
Manufacturing
ERP analysis, quality manual search, supply-chain reports, staff training.
02
Finance and insurance
Product terms Q&A, compliance search, agent training.
03
Healthcare
Drug and guideline knowledge bases, compliance training, patient FAQs.
04
Trade and distribution
Product knowledge, customer service bots, reports, GEO.
05
Education
Course generation, exams, progress tracking, certificates.
06
Technology
Technical docs Q&A, runbook search, onboarding exams.

FAQ

Does our data go to the public internet?
No. The platform runs on your own servers; data stays inside your network and actions are logged.
What server do we need?
Recommended 4 CPU cores, 16 GB RAM, 100 GB SSD; minimum 2 cores, 8 GB, 50 GB. Local models need a GPU.
What service is included?
Installation, regular updates, online support and troubleshooting. Custom development, hardware and on-site work are quoted separately.

Book a first conversation

We learn where you are and what you need, then propose a plan.

Book a consultation

Send us a message

Contact

Choose a topic and leave your details. We reply within one to two working days. You can also call our hotline or add us on WeCom.

Address
Room 1707, Building T2, Yuehai City, Luohu District, Shenzhen
WeCom QR code

WeCom

Scan with WeChat to add us

Topic