Skip to content
LeapWise
Menu
中文
All solutions

Office & knowledge

The easiest place to start, in any industry. People use it straight from WeCom or Feishu, results show within a month or two, and it extends into business systems from there.

Which receipts do I need?12PolicyProcessillustrative

Featured solution

Knowledge-base AI worker

Every employee finds the answer with one question in chat — and every sentence links back to its source

How it works

  1. 1

    Connect sources

    File shares, drawing vault, scans

  2. 2

    Sync permissions

    Follows the source system, per person

  3. 3

    Ask

    WeCom, Feishu, DingTalk

  4. 4

    Cited answer

    Each sentence cited and linked

  5. 5

    Feedback loop

    Stored only after confirmation

What you get

  • New hires no longer depend on asking veterans
  • Answers can be checked, so made-up answers get caught
  • Material stays inside; permissions are respected

What we deliver

  • Private knowledge-base system
  • AI worker inside your chat tools
  • Evaluation set built from your real material
  • Acceptance report and operations manual

Featured solution

AI office assistant

An AI office workspace for everyone, with usage, cost and content safety under control

How it works

  1. 1

    Employee request

    Minutes, policies, documents, slides

  2. 2

    Content filtering

    Sensitive terms and personal data

  3. 3

    Model gateway

    Department quotas, cost allocation

  4. 4

    Delivered

    Editable documents

  5. 5

    Audit

    Traceable end to end

What you get

  • Writing work gets noticeably faster
  • No more private use of outside AI; leak risk under control
  • Clear usage and spend for every department

What we deliver

  • Private office workspace
  • Office template library
  • Model gateway and quota policy
  • Audit and operations reports

AI office assistant

Deliverable

One workspace for meeting minutes, policies, official documents and slides — with department quotas, sensitive-data filtering and a full audit trail.

The problem
Writing takes too long and formats drift; staff using outside AI risk leaks; usage and cost are out of control.
Who it's for
Group HQs, state-owned enterprises, public institutions, corporate functions
Delivery
Private deployment, delivered in WeCom, Feishu, DingTalk or on the web
What it does
  • Draft minutes, policies, documents and slides in one step from built-in templates
  • One model exit: sensitive-term and personal-data filtering, department quotas, cost allocation
  • Encrypted storage for files and messages, long-term audit logs
  • Works in WeCom, with intranet allow-listing

WeCom office agent

Deliverable

An AI worker that lives in WeCom with role-specific skills: weekly reports, SOPs, official documents, 8D reports and scheduled reminders.

The problem
Front-line and support staff spend hours a day writing, searching and chasing progress.
Who it's for
Company-wide office work at small and mid-sized manufacturers
Delivery
Private deployment, delivered in WeCom, Feishu, DingTalk or on the web
What it does
  • Skill packs per role: admin, purchasing, quality, production
  • Weekly reports, SOPs, documents and 8D reports drafted from one sentence
  • Look up drawings, policies and ledgers right in the chat
  • Reminders and daily summaries pushed to groups automatically

Knowledge-base AI worker

Deliverable

Ask in Feishu or WeCom; answers come from your own process documents, scans, drawings and ledgers, with a source on every sentence.

The problem
Material is scattered and the latest version is hard to find; new hires depend on veterans; general AI doesn't know your standards and makes things up.
Who it's for
Engineering, process and quality teams; document-heavy manufacturing and engineering firms
Delivery
Private deployment, delivered in WeCom, Feishu, DingTalk or on the web
What it does
  • Multimodal indexing: Office, PDF, scans, CAD drawings, ledgers
  • Semantic search plus direct drawing/part-number lookup; every sentence cites and links back to its source
  • Permissions follow the source system, filtered per person
  • Indexes through file-encryption systems; data never leaves the company

Policy and procedures assistant

Deliverable

Staff ask which policy applies, who to ask and what to do; answers quote the policy currently in force, with version and effective date.

The problem
Too many policies and versions — nobody knows which is valid, and the same questions keep landing on support teams.
Who it's for
Corporate groups, state-owned enterprises, financial institutions, government and public bodies
Delivery
Private deployment, delivered in WeCom, Feishu, DingTalk or on the web
What it does
  • Full policy lifecycle: draft, publish, in force, superseded or withdrawn
  • Searches only clauses in force that the employee may see
  • Answers structured as conditions, steps and required documents
  • Links straight to the right service; hands off to a person when it can't answer

Document studio

Deliverable

One dataset produces a Word report, an Excel workbook with real formulas and a slide deck — with every figure matching across all three.

The problem
The same data has to become three documents; copy-paste errors leave the numbers inconsistent.
Who it's for
Document-heavy teams such as business analysis, finance and general management
Delivery
Private deployment, delivered in WeCom, Feishu, DingTalk or on the web
What it does
  • Uploads are screened first: macros, external links and embedded objects are rejected
  • Facts come from tables; text sources fill gaps and are checked word by word
  • Derived metrics are computed by code — the model never invents numbers
  • Editable Word, Excel and PowerPoint output per template and audience

Meeting minutes and follow-up

Deliverable

Turns meeting transcripts into confirmed minutes, decisions and owned action items — then follows them through.

The problem
Nobody takes notes, nobody follows up, and decisions never get done.
Who it's for
Management meetings, project stand-ups, cross-team coordination
Delivery
Private deployment, delivered in WeCom, Feishu, DingTalk or on the web
What it does
  • Imports transcripts and matches speakers to the attendee list
  • Extracts decisions and actions, converting deadlines to dates
  • Takes effect only after both the chair and the owner confirm
  • Writes to the task system with due-date reminders and nudges

Contract and procurement review

Deliverable

Reviews contracts against your own clause standards and cross-checks requisitions, quotes and contracts; proposed edits go to the owner for confirmation.

The problem
High contract volume, few lawyers; requisitions, quotes and contracts don't line up.
Who it's for
Legal, purchasing and tender management
Delivery
Private deployment, delivered in WeCom, Feishu, DingTalk or on the web
What it does
  • Clauses split and classified: payment, acceptance, breach, IP, confidentiality and more
  • Every finding is checked against the original text; amounts are computed by code
  • Three-way match of requisition, quote and contract
  • Only human-approved edits are applied; nothing is signed or sent automatically

Ask-your-numbers analytics

Deliverable

Ask about business metrics in plain language; answers use approved metric definitions and reproducible queries.

The problem
Executives wait for reports to get a number, and each department defines the same metric differently.
Who it's for
Executives, business analysis and finance teams
Delivery
Private deployment, delivered in WeCom, Feishu, DingTalk or on the web
What it does
  • Metrics published after two-person review, with one definition
  • The model only plans the query; execution is fully deterministic
  • Organisation scope injected server-side, backed by row-level security
  • Saved questions and scheduled reports

Start with a half-day workshop

Your business owner and IT lead join us to pin down 1–3 candidate use cases and their success metrics.

Book a use-case workshop

  1. 1

    Use-case workshop

    Half a day, on site or online

    • Business owner and IT lead attend
    • Map 1–3 candidate use cases and the state of the data
    • Output: use-case list and draft success metrics

  2. 2

    PoC proposal

    Within a week of the workshop

    • Data and system boundaries, evaluation-set design
    • Milestones, commitments on both sides, acceptance bar
    • Output: a PoC proposal ready to sign

  3. 3

    Run the PoC

    2–4 weeks

    • Your real material, in a private or demo environment
    • Weekly check-ins with visible metrics
    • Output: evaluation report and launch plan