About us

A company that puts AI into specific business processes

We do not build foundation models and we do not sell reports. We build the engineering that fits AI into real workflows, then use that engineering to run four product lines independently.

What we do, what we do not

We do

  • Fit AI into existing workflows without asking you to replace systems first
  • Build reusable engineering: model access, data pipelines, deployment and operations
  • Run four product lines independently, each with its own entry
  • State the boundary before discussing scope

We do not

  • Sell reports with no implementation
  • Promise conversion rates, closing rates or growth numbers
  • Use customer details for anything beyond the current enquiry
  • Publish case studies or customer lists without documented consent
Engineering base

Four capabilities on one base

The four product lines share the same underlying capability, not the same interface.

  • Channel layer

    Only connects: short video, community groups, social feeds and search, without touching channel settings.

  • Intelligence layer

    Only judges: intent reading, scoring, conversation and attribution, returned as structured output.

  • Handover layer

    Only passes on: structured results go to a specific person, with a record kept.

  • Engineering base

    Model access, data pipelines, deployment and operations, shared by all four lines.

System layers: channel, intelligence, handover Channels above, judgement in the middle, sales action below, all joined by one data structure. SYSTEM LAYERS L1 · Channel layer Short video · groups · social · search Content and budget land here 01 L2 · Intelligence layer Intent · scoring · conversation · attribution Traffic becomes qualified leads here 02 L3 · Handover layer Sales workspace · talking points · cadence People take over; AI records and reminds 03 No manual data moving: L2 output is the input for L3.
Diagram: layers exchange one data structure, so no data is moved by hand.

Four product lines

One base, four directions.

  • Jiutian XuanhuaAIGC short-drama creation platform · WeChat mini programWeChat mini program
  • Muxue AI lead generationAI lead generation · customer operations SaaSwww.mxly.plus
  • StarHub international settlement systemCross-border settlement and finance systemstarhub.mxly.plus
  • Agent software automation platformAI agent development and delivery automationxq.mxly.plus

Timeline

Only confirmed facts are listed. Unconfirmed milestones are left out rather than guessed.

  1. 2015.12

    Company founded in Shenzhen to build software and applied AI.

  2. To be confirmed

    Further milestones are pending from the founder.

Verifiable facts

Each item below can be checked through public records. Unconfirmed items stay blank rather than filled in.

ItemDetail
Registered name (Chinese)深圳市丹凤玛莉智能科技有限公司
Registered name (English)Shenzhen Danfeng Mali Intelligent Technology Co., Ltd.
FoundedDecember 2015
Registered locationShenzhen, Guangdong, China
Business focusSoftware and applied artificial intelligence
Product linesFour (three subdomains + one WeChat mini program)
ICP filing粤ICP备2023059377号
Office addressFloor 1, Building 5, Liantang Jintian Industrial Zone, No. 128 Aoxia Village, Luohu District, Shenzhen, Guangdong, China

Check first, then talk

Verify each item above, or simply ask us.

Start with the scenario

Tell us your industry and how to reach you. We check whether the work fits before discussing scope.

Fields marked * are required

Used to address you, nothing else

We read the industry first to judge fit

Phone or email only; WeChat ID, QQ number and landlines are not accepted

10 to 500 characters, at most 2 links

We reply within one business day. Your details are used for this enquiry only.

We reply within one business day. Your details are used for this enquiry only.

Talk to us