ShenZhen Nebula Rover Technology Limited·深圳市星云智航科技有限公司Hotline400-990-0246Emailperry.xu@airover.net
AIPaddle · Enterprise AI platform

Turn the AI in your people's hands
into AI your company owns

AIPaddle is an enterprise AI platform that lives inside WeCom: build AI in one sentence, run each project from one group, connect to SAP and other ERPs, and choose SaaS or on-premises. Nebula Rover builds it and helps you roll it out.

  • No learning curve
  • AI in a sentence
  • Editable by people
  • Company-owned
  • AI project manager
  • Governed, deployable
Sample project · Project group WeCom · illustration
Mon 08:00
PMO
PMO assistant
Daily check: T-004 lease ledger review is due tomorrow @Asset manager; T-007 is 2 working days overdue @Renovation lead please update the status.Basis: Project Management Policy · escalate to PM after 5 working days overdue
PM
Project manager
@PMO assistant Assign to 〈Leasing lead〉: anchor-tenant rent proposal by Friday
PMO
PMO assistant
Created task T-012, due Fri 9 Oct, added to the Project Execution Plan. @Leasing lead
Illustrative conversation; 〈〉 marks a member's name ·
What leaders keep asking

Everyone uses AI, yet the company is no stronger

The problemHow AIPaddle answers it
Know-how walks out the door

The best prompts and workflows live in personal laptops and accounts and leave with people

AI is built on the company platform and owned by the company
Everyone reinvents the wheel

The same weekly reports, news briefs and contract checks are rebuilt by every department

Build once, approve, reuse everywhere
New systems don't stick

Another app and another login, and frontline staff won't use it

It runs inside WeCom, nothing to install
Projects depend on chasing

Assigning, chasing, risks and reports depend on the PM's memory; policies go unenforced

The PMO Agent runs projects by your policies
No control, no numbers

Nobody can see who built what, which data it read or what it cost

Permissions, approvals, activity and usage on one screen
Inside the product

Not a concept, a system in daily use

Build in a sentence
The assistant lists a build plan first: which objects, which data; nothing is created until you confirm.

Real product screenshots from a customer's production environment, customer details removed

SAP / ERP integration

AI that can read your ERP
and respects its permissions

APIs only, never direct database access; read first, write-back only with approval; encrypted credentials and full audit. Five channels cover reads, events, inbound calls, write-back and MCP.

  • OData / API Management
  • Event Mesh events
  • Fiori / BTP calls AI
  • Controlled BAPI write-back
  • OData MCP
  • IAS / IPS identity
See the SAP architecture in motion →
WeComWeb
AIPaddleAPI · MCP · WebhookIAM · Audit
SAPBTP · API MgmtS/4HANA · ECC
In production

A real estate asset manager: one month from sign-up to AI-run projects

  1. Day 1Project created

    WeDrive space, folders and trackers created automatically

  2. Day 10Workflow in a sentence

    Daily news brief pushed automatically

  3. Day 16Agents approved

    Asset review assistant and others granted to staff

  4. Day 21Assignments and risks in groups

    PMO checks daily and pushes risk follow-ups

  5. Day 22Short videos live

    Asset library and daily short video

77

AI objects (20 agents · 14 workflows · 16 skills · 6 digital employees, etc.)

168

documents in 12 knowledge bases, shared by person or department

5

scheduled jobs running daily

98.4%

call success rate

Figures from the customer's production environment (5 Oct 2026); the customer is not named by agreement. All case studies →

Getting started

Start with one project and one department, see results in two weeks

  1. Week 1

    Connect

    Open SaaS or deploy on-premises, sync your WeCom org, bind one project group.

  2. Week 2

    Prove it

    Put the PMO on one key project; one department builds 3 high-frequency AIs with the assistant.

  3. Month 1

    Scale

    Approve and publish what works for everyone; connect SAP or other ERPs when needed.

  4. Ongoing

    Decide with data

    Use usage, cost and acceptance rates to choose what to build next.

How fast a full rollout goes depends on your team's involvement and the complexity of your business and organization; we give you a concrete plan during the assessment.

AI organization rollout services

We help companies go from "individuals use AI" to "the organization runs on AI"

The product is only half of it. Our delivery team works in four stages, each with clear deliverables.

  1. 01

    Assess

    Interviews and inventory: which roles, processes and documents to start with; current AI tools, ERP and data risks.

    Assessment report · priority use cases
  2. 02

    Set up the foundation

    AIPaddle SaaS or on-premises; WeCom, WeDrive, directory, ERP, enterprise identity and model providers connected; roles and permissions.

    Working tenant · integration overview · permission matrix
  3. 03

    Build assets

    Knowledge bases, skills, agents, workflows and digital employees for the priority use cases; project templates and PMO rules.

    First AI objects published and authorized
  4. 04

    Operate and measure

    Training and acceptance testing; iterate on usage, acceptance rate and KPIs.

    Training materials · acceptance records · operating reports
Deployment

One codebase, two ways to deploy

On-premises is not a separate product: platform operations features are removed at build time, and your servers hold only compiled output, never source code.

SaaS

Ready to use

  • AIPaddle cloud service, available as soon as it is opened
  • Row-level multi-tenant isolation; companies never see each other's data
  • Merge → full test suite → automatic deploy, continuously updated
  • Use platform models or bring your own model provider
On-premises

On your own servers

  • A dedicated database that holds only your company
  • No source code delivered, enforced by an allow-list gate
  • Models called through a gateway; real keys never touch your servers
  • Migrations run locally on your server, with daily backups
Book a demo

Start with one project group

We first get the PMO Agent running in one of your real project groups, and you can usually see results within two weeks. How fast a full rollout goes depends on your team's involvement and on the complexity of your business and organization; we give you a concrete plan during the assessment.

  • A 30-minute online demo built around your industry
  • How we would connect to your ERP and WeCom
  • A proposed pilot scope and plan

Hotline: 400-990-0246

Or write to us directly: perry.xu@airover.net