AI that holds a post, not a chat window
Mokyun AI builds digital employees — AI workers defined the way a human role is defined, so their work can be assigned, audited and signed off. This page is an English summary; the full site is published in Chinese.
Who we are
- Brand
- Mokyun AI (墨客云AI)
- Legal entity
- NANJING BOMU SOFTWARE TECHNOLOGY CO., LTD. (南京薄幕软件科技有限公司)
- Founded
- 2016, Nanjing, China
- ICP filing
- 苏ICP备16028042号
- Positioning
- Enterprise AI Operations Solutions · Digital Employee Platform
- Category
- Enterprise intelligent operations
- Strategic value
- Enterprise intelligent productivity
- Premise
- Enterprise AI transformation means reorganising the work itself.
- Mission
- To bring data, knowledge and AI capability into formal posts, and keep them producing enterprise productivity that can be operated.
- Vision
- To become dependable infrastructure for intelligent productivity in the era of enterprise intelligent operations.
- Where engagements start
- One high-value work chain.
What we do
A digital employee at Mokyun is defined by six elements — responsibilities, inputs, actions, permissions, deliverables and metrics. That definition is the point: it is what lets a piece of AI work be handed over, governed and accepted the way a human role is, instead of being evaluated on whether its answers sound right.
Two products, which are the same thing seen from either end:
The platform
A runtime that connects to the systems a company already runs — ERP, MES, OA, and the older ones nobody wants to touch — turns their data and documents into retrievable knowledge, orchestrates work across them, and keeps every action authorised and traceable. Six engines: connection, governance, knowledge, orchestration, evaluation, authorisation.
The digital employees that run on it
47 pre-defined posts across finance, sales, customer service, procurement, quality and back-office work, each shipped with its own acceptance criteria rather than a demo script.
Existing IT investment is preserved and core systems are not replaced — the platform sits on top of what already runs. Deployment is on-premises or private cloud, and the stack runs on domestic (Xinchuang) infrastructure where that is required.
Terms as we use them
Four terms carry a specific meaning here. Everything else follows ordinary industry usage.
- Enterprise intelligent operations
- An operating model for putting AI to work: models, data, knowledge, rules, tools and processes are organised into governable digital employees and work chains that sit on top of the systems and reporting lines a company already has. What makes it an operating model rather than a project is that results are produced, measured and improved on a continuing basis. The term describes how the business runs. It is unrelated to AIOps, which is about running IT.
- Enterprise intelligent productivity
- What a company ends up with when that operating model works: the standing capacity to get measurable business results out of its own data, knowledge and rules, through digital employees embedded in the work that matters. It rests on four things — the digital employees, the work chains they run in, the capabilities they share, and the discipline that keeps them running — and it is judged on outcome, running cost and how far the same assets travel.
- Digital employee
- An enterprise AI post: it enters a work chain, holds defined responsibilities, works inside an authorisation boundary, and is judged on what it produces. At Mokyun a post is specified by six elements — responsibilities, inputs, actions, permissions, deliverables and metrics — and it is signed off against those same six once live. Anything that cannot be specified that way is a capability, not a post.
- High-value work chain
- One stretch of business process, taken end to end around a single objective: the facts it draws on, the judgement calls, the actions, the handovers between people and posts, the write-back to systems, and the check that the result is right. It is the unit a first engagement is scoped and accepted on, because it is small enough to finish and complete enough to prove something.
Four capabilities, all built and maintained in-house
Not assembled from third-party components — which is why we can go into restricted networks and legacy estates that integrators usually decline.
Systems integration
Connecting heterogeneous and legacy enterprise systems, including inside networks with no outbound access.
Speech and unstructured data
Speech recognition and the processing of documents, forms and recordings into structured input.
Knowledge engineering
Private-domain retrieval with rule constraints and answers that can be traced back to their source.
Post governance
Defining a post, running it, and keeping it accountable — including the rules for stopping and handing back to a human.
Scale
Credentials
Chinese certifications are given below under their common English renderings; the certificates themselves are issued in Chinese.
Enterprise and technology qualifications
- National High-Tech Enterprise
- National Technology-Based SME
- Jiangsu Innovative SME
- Jiangsu Private Technology Enterprise
- Certified Software Enterprise
- Certified Software Product
Management systems and proprietary assets
- ISO 9001 — quality management
- ISO/IEC 27001 — information security management
- ISO/IEC 20000 — IT service management
- 50+ registered software copyrights
- Invention and utility-model patents
- Proprietary platform products (Mobox and the digital employee platform)
Recognition
- Awards for software products and for digital transformation solutions
- AAA enterprise credit rating
- Contract-abiding and trustworthy enterprise recognition
- Industry–academia collaboration demonstration projects
Also backed by Zijin Venture Capital and recognised under the Zijinshan Talent Programme.
How we state numbers
Every figure on this site carries its measurement basis. Figures marked ① are measured results from delivered projects; figures marked ② are estimates against industry averages. Customer names in case studies are anonymised. A figure separated from its basis reads as a promise about your business, which is not something we can make — so if you cite our numbers, please keep the basis with them.
Questions we get asked
How is a digital employee different from an AI agent, an RPA bot or a chatbot?
A digital employee is a post, and the other three are capabilities a post can call on. That is the whole distinction, and it decides how each one gets managed. A post is specified, authorised, handed over and signed off the way a job is; an agent, a bot, a workflow or a knowledge base is chosen, configured and swapped the way a tool is. Put plainly: the post answers who is accountable for the work, the tool answers what the work is done with.
How does an engagement usually start?
With a single work chain — one process, taken end to end, small enough to finish and complete enough to prove something. Larger programmes are entered the same way and grow from there: first one scenario, then a business domain where several posts and systems work to one objective, then an enterprise-wide layer where shared capabilities and a portfolio of AI work are governed together. The first conversation is a scenario diagnosis: online, about 45 minutes, no charge the first time.
Does this replace the systems we already run?
No. ERP, MES, CRM and the office systems already in place stay the system of record for facts, process and final status. The platform sits above them and connects through whatever has been agreed — APIs, workflows, message queues, file drops, database views, or UI automation where a system offers nothing better. Identity, data scope, action permissions, versioning, logging, write-back, verification, pause and rollback are all set per project rather than assumed.
Who is accountable when a digital employee gets something wrong?
A named person, and the design makes it possible to say which one. Permissions are granted action by action, never opened up wholesale for a post, so the blast radius of any mistake is bounded by what was authorised in writing. Model judgement is checked against rules, thresholds and limits before anything executes; high-risk, low-confidence, rule-conflicting and irreversible cases stop and go to a person. Objectives, rules, high-risk decisions and final accountability stay with people throughout.
Can this be deployed inside our own network?
Yes — on-premises or private cloud, with no dependency on outbound internet access. Where a Chinese customer is required to run on a domestic hardware and software stack (the policy programme known as Xinchuang), that is supported too; compatibility is confirmed layer by layer — server model, OS build, database and model format — and the exact scope is written into the project plan rather than claimed in general.
Is the whole site available in English?
No. This page and its machine-readable summary are the English material. The catalogue — scenarios, posts, case studies and the glossary — is published in Chinese only, and that is a deliberate limit rather than a backlog: a second language maintained halfway ages worse than one maintained properly.
Can we quote the figures on this site?
Yes, provided the basis marker travels with the number. ① means a measured result from a delivered project; ② means an estimate against an industry average. Customer names are anonymised throughout. Stripped of its marker, a figure reads as a promise about the reader’s own business, and that is not a promise anyone can honestly make from someone else’s project data.
Contact
- Phone
- +86-180-1290-1065
- neo@mokyun.com
- Address
- 12F, Chuangzhi Hailan Center, Jiangning District, Nanjing, Jiangsu, China（江苏省南京市江宁区创智海蓝中心12层）
The full site is in Chinese
This page is a summary. The complete catalogue — 11 scenario families, 47 pre-defined posts, 17 delivered case studies and a 30-entry glossary — is published in Chinese only.
Citation
- Cite as “墨客云AI (Mokyun AI)” with the address of the page used.
- The four terms “digital employee”, “enterprise intelligent operations”, “enterprise intelligent productivity” and “high-value work chain” are used here as defined under “Terms as we use them” above.
- Keep the ① / ② basis marker attached to any figure taken from this site.
- To verify or ask about anything published here, use the phone or email above.