HomeProductUse casesPricingAboutHelp center
Product

A sales knowledge system light enough to actually use

From WorkLog capture to AI prediction, every customer touchpoint and every document grows on the same customer knowledge network. See what Youdian is made of.

Product overview

One data path from record to prediction

Every touchpoint and every document travels the same path: input, AI processing, graph data, application—then settles into reusable customer assets that let you plan ahead.

INPUT

Input

Voice WorkLogText inputPPT / PDF uploadOffline cache
AI

AI processing

Transcribe and summarizeEntity recognitionAutomatic backlinksProfiles and prediction
GRAPH

Graph data

Contacts / companiesProjects / knowledge cardsTwo-way linksPermission filtering
APP

Application

ProfilesProject boardAI Q&APrediction and reminders
Capture

Say a sentence, and a customer visit files itself

Voice, text, and files all go in, with offline cache and weak-signal retry. AI extracts events, people, companies, and projects, then notifies you when it is done—capture never breaks your rhythm again.

  • Press and talk, filed in 30 seconds
  • Auto-extracts events / people / companies / projects
  • Offline cache and weak-signal retry lose no records
  • Push notification when AI processing is done
9:41
Recording client visit
00:27
Live transcript

Visited United Nutrition Tech today and met Lin Lan. Talked about the HN-021 purchase project. They care about Q3 supply; need to send the stability report next week…

AI recognized 3 entities
Auto-linked to the existing customer graph
Lin Lan
Contact · Head of Procurement
98%
United Nutrition Tech
Company · Existing profile
99%
HN-021 purchase project
Project · Quoting
94%
Customer profile

The more you record, the fuller the profile

Company profiles, contact profiles, and project profiles grow automatically from your WorkLog: objective facts are merged and deduplicated, subjective judgments keep their source, and history is never overwritten.

Company profile

Company Profile
  • Overview and objective facts
  • Interaction timeline
  • Document center
  • Related network
  • AI overall judgment

Contact profile

Contact Profile
  • Role and decision weight
  • Interaction history
  • Related projects
  • Relationship network
  • AI overall judgment

Project profile

Project Profile
  • Stage state machine
  • Project timeline
  • Quotes and documents
  • Key contacts
  • AI overall judgment
Company Profile
Active client

United Nutrition Tech

Functional food ingredientsShanghai · 200–500 staffLead: Lin Lan
Objective info
Linked projects3 in progress
Total quoted¥ 2.4M
Last touch2026-06-08
Touchpoints47
Key contacts
L
Lin Lan
Head of Procurement · Decision-maker
Z
Zhou Mingzhe
R&D Lead · Influencer
C
Chen Siyuan
QC Manager · User
AI synthesis

The client is focused on Q3 supply stability and price lock-in. R&D has passed sample testing; the deal is stuck at the annual budget approval by the Head of Procurement.

Source: 3 WorkLogs · 2 meeting notes · 1 quote
Project management

See the path a project is taking from customer touchpoints

Project stages are recognized from records and history settles automatically. Sales reps can correct by hand, managers view the whole team by stage—even the reasons a deal was lost are kept.

  • Stage state machine advances automatically
  • Kanban board by stage
  • Project owner and next action
  • Loss reasons archived for review
Pipeline
This quarter · 20 deals
First contact4
Aurora Bio
¥ 480K
G6/12
Needs confirmed3
United Nutrition Tech
¥ 1.2M
L6/15
Quoting5
Haiyue Foods
¥ 860K
W6/10
Mingyuan Health
¥ 540K
L6/18
Negotiation2
Greenfield Channels
¥ 2.1M
G6/20
Closed6
Zhonghe Nutrition
¥ 760K
WSigned
Knowledge graph

Not just a customer list, but a relationship network

Contacts, companies, projects, and knowledge cards are linked by backlinks into one relationship network shared by the team. See at a glance who knows whom and who is blocking whom; open any node to reach its profile page.

  • Six entity types linked to each other
  • One-hop relationship network visualization
  • Click a node to jump to its profile page
  • A secure view after permission filtering
Knowledge Graph
One-hop links
United Nutrition TechCompany
Lin LanDecision-maker
Zhou MingzheR&D
HN-021 PurchaseProject
Stability reportDocument
GB 28050 RegulationRegulation
AI prediction

Once enough builds up, AI starts to plan ahead for you

Once profiles and history reach a tipping point, Youdian does more than answer—it starts to plan ahead: what to do next, which deal is stalling, which window is best to follow up. Every prediction traces back to the original record.

  • Next-action suggestions with reasons
  • Early warning for stalls and churn risk
  • Best follow-up timing hints
  • Predictions keep recalibrating with new records
AI predictions · Next Steps
Predictions from 312 records and profilesUpdated daily
Suggest revisiting Lin Lan this week to push budget approval
She mentioned approvals usually close by mid-June · based on 3 records
View basis
Haiyue Foods deal risks stalling
No reply for 14 days after quoting; similar past deals often stall here
Follow up
Mingyuan Health is entering a reorder window
Historical cycle is about 90 days; this week is a good time to quote
Prepare quote
Growing knowledge base

You use it, AI grows it

Using it is simple: upload PPTs and PDFs, ask questions as you go. Growing it is on the AI: split into knowledge cards, backlink to customers and projects, update old cards with new material, run regular checkups. The originals are always kept and always traceable—the knowledge base gets both richer and fresher the more you use it.

Knowledge base · Knowledge Cards
Health check passed · today128 knowledge cards
Upload PPT / PDFAuto-split into cardsCited in AI Q&A Answers settle into new cards
Product doc
HN-021 spec sheet
  • Purity ≥ 99.2% · water-soluble
  • Q3 capacity 40T / month
From PDF · split into 12 new cards, 3 updated
AI-distilled
HN-021 vs competitors
  • 18% higher solubility
  • 6% higher price, better stability
Distilled from one AI Q&A
Regulation
GB 28050 labeling essentials
  • Nutrition-claim threshold reference
  • Label review checklist
From PPT · split into 6 cards

New material grows into old knowledge

A single PPT is split into new cards and updates the related old ones along the way—when a quote changes or a model is upgraded, everything follows.

Answers cite sources, traceable to the original

Q&A cites specific cards, each card records where it came from, and you can trace all the way back to the original file.

AI runs regular checkups

Conflicting numbers, expired quotes, orphan cards no one cites—AI finds them and reminds you to handle them.

AI Q&A

Have a question? Just ask the knowledge base

What was last year's quote? Where is the project stuck? Who is in the decision chain? Ask in one sentence—AI answers from the team's real records and knowledge cards, every answer cites its source, and a good answer settles into a new card with one click.

  • Ask in natural language, answers in seconds
  • Every answer carries a source citation
  • One click turns an answer into a knowledge card
  • The profile's overall judgment keeps updating
Ask Youdian
What prices did we discuss with United Nutrition Tech last year? Where is the deal stuck now?
Youdian AI

Two quotes last year: ¥1.4M in May, ¥1.2M in November. The HN-021 deal is currently in Quoting, R&D has passed sample testing, and the blocker is the annual budget approval by Lin Lan, Head of Procurement.

Quote record · 2025-11-14
WorkLog · Zhou Mingzhe visit · 2026-05-30
Meeting notes · Quarterly review · 2026-06-02
Keep asking — answers settle into the customer knowledge base…
Security and permissions

Team sharing, with room for personal judgment

Youdian separates WorkLog originals, AI summaries, objective facts, and subjective judgments—different roles see different layers, so sharing stays sharing and boundaries stay boundaries.

CapabilitySalesLeaderOwner
Record WorkLogOwnOwnAll
View own originals
View reports' originals
View reports' summaries
View shared entity activity
View team activityOwn groupWhole company
Manage team structureOwn groupWhole company

Field-level encryption

WorkLog originals are encrypted at the field level, and files are stored with encryption at rest.

Subjective info isolation

Objective facts and subjective judgments are separated, visible in layers by role.

Access log audit

Sensitive actions leave a full trail that the owner can audit.

Offboarding data handling

When a member leaves, customer assets stay with the company for a smooth handover.

Roadmap

From lightweight capture to a full sales knowledge system

Phase 1
In progress · MVP

Lightweight capture

  • Voice / text capture
  • WorkLog summaries
  • Entity recognition and automatic backlinks
  • Basic customer profiles
Phase 2
Knowledge capture

Documents and Q&A

  • PPT / PDF upload
  • Automatic knowledge cards
  • AI Q&A with citations
  • Answers settle into new cards
Phase 3
Team collaboration

Multi-role collaboration

  • Multi-role permissions
  • Project board
  • Admin console
  • Lint reports
Phase 4
Compounding loop

Profiles and prediction

  • Deeper company / customer profiles
  • AI prediction and action suggestions
  • Proactive reminders
  • Organizational knowledge assets

Make customer relationships a company asset that compounds

Start with a single WorkLog, so every customer touchpoint is captured, connected, and usable.