Tiana Tiana

乐于思考、喜欢探索的产品经理一枚，期待转向AI产品经理。 过往工作涉及了B端平台设计、渠道运营、LLM数字助理与Workflow自动化，目前正在探索AI Agent如何解决真正的问题。

Based in Beijing

View AI projects → View AI projects → Contact Me → Contact Me →
01

About Me

Core Positioning AI Product Manager

B2B workflow insight × Agent implementation × Operations efficiency

02

Featured AI Projects

LLM Agent Open Claw + mimo Scenario Definition 5 Skills Designed Prompt & Feedback Iteration

Digital Assistant "Xia" | Operations AI Agent built on Open Claw

Built a 0-1 LLM Agent for store incentive operations, addressing scattered knowledge, repeated high-frequency questions, and standard actions that previously relied on manual experience.

High-frequency question Knowledge retrieval Skill judgment Process guidance Standard response
Agentic Workflow Claude Code + GPT-5.5 4 MCP Tools Workflow Engine 95% Efficiency Gain

Calculation Engine Acceptance Agent | Turning manual checks into Agentic Workflow

Encapsulated policy queries, price checks, system result retrieval, rule matching, and conclusion generation into a script-based AI acceptance workflow.

Input policy ID Query policy / price Fetch system results Match rules Output conclusion
03

B2B Platform Projects

Calculation Engine Rule Modeling Project Owner Scenario Standardization

Store Incentive Restructuring

作为项目主Owner，主导门店激励重构，搭建计算引擎，协调7方角色，涉及3个系统，覆盖政策创建、下发、实时计算/公示、对账和结算全链路。

IR policy types streamlined from 12 to 8 Reconciliation data refined to IMEI level Online calculation rate 40% → 72% Store participation 48% → 58%
Store Incentive screenshot 1 Store Incentive screenshot 2 Store Incentive screenshot 3
Project guide card
Project in one sentence

Move store incentives out of RMS and rebuild them as a sales-side calculation engine.

RMS dependency Backend capability was tied to the retail system.
Fragmented rules Each policy type had its own calculation path.
Weak closed loop Policy, calculation, reconciliation, and settlement were not fully unified.
01 Policy creation Sales configures standardized IR incentive rules.
02 Real-time calculation The engine calculates levels, ladders, and incentive amounts.
03 Publication Results are sent to Distributor App for store visibility.
04 Reconciliation & settlement IMEI-level data supports reconciliation, adjustment, and settlement.
Independent Retailer（IR） Retailer receiving incentive results.
IMEI Device-level grain for calculation and reconciliation.
SO / ST Sell-out and sell-through data used by different policy rules.
RRP/STP Suggested retail price used by price-related checks.
Policy types 12 → 8
Calculation online rate 40% → 72%
Store participation 48% → 58%
Reconciliation grain IMEI
0-1 Project Owner Cross-system Collaboration End-to-end Retailer Management

Store Incentive M+1 Settlement

Designed an end-to-end online reconciliation and batch settlement flow across Distributor App and CRM, covering Dealer approval/rejection, dispute handling, batch merging, status sync, and settlement disclosure.

Pilot covered 1 country and 1,773 merchants Processed 7,092 bills Settlement cycle 104 days → 22 days NSS 63% → 73%
试点国家用户反馈
Project guide card
Project in one sentence

Turn IR incentive reconciliation and settlement into an online, transparent, batch-based M+1 process.

Manual trigger Reconciliation and settlement relied on people to start and approve.
No batch settlement All downstream IR data had to be ready before one unified settlement.
Opaque reconciliation MD and IR could not both see clear incentive details online.
01 Data ready Identify IR incentive data that is ready after the activity ends.
02 Online reconciliation MD confirms, rejects, or handles disputes with visible IR details.
03 Batch merge Ready IR bills can enter batches first instead of waiting for all IRs.
04 Settlement disclosure Settlement status and results are synchronized and disclosed online.
IR Independent Retailer receiving incentive settlement.
Agent Master Dealer confirms reconciliation and uploads invoices.
Finance Reviews settlement progress and payment status.
IR bill The bill grain used for reconciliation and batch settlement.
Pilot scope 1 country / 1,773 merchants
Bills processed 7,092
Settlement cycle 104 days → 22 days
NSS 63% → 73%
04

Story

2025.11 - 至今

AI 实践

自主探索 Agent 运营提效与 AI Coding

2025.07

主 Own 门店激励

主导了系统重构、M+1 结算等项目

2024.06

入职小米

担任 B 端产品经理，All in 分销通 APP

2023.07

美团实习

担任数据产品经理，参与数据治理

2022.10

"互联网+"创新创业大赛

无人机机群智能管控 - 团体银奖

2021.09

研究生学习

发表 1 篇 SSCI 论文、1 篇 EI 会议论文、3 篇发明专利

2021.01

入职字节跳动

从事 AI 语音相关的数据标注

2020.06

入职东方海外物流

主跟佳能进口海运单

2016.09

本科学习

担任班长，主修运筹学、概率论和计算机基础等

05

AI 资讯日报

日一二三四五六

Loading...

06

Resume

Download my PDF resume for full project details.

Download PDF Resume
07

Contact

Open to AI Product Manager opportunities in Beijing.

1278020254@qq.com TYJ6DFQ Back to Top