dingtalk-workflow-business-advisor

AI Business Advisor Team (dingtalk-workflow-business-advisor)

8 AI expert roles automatically analyze business data every night, summarize and deduplicate, then send a numbered list to DingTalk group chat/DING, with support for replying with a number to expand detailed analysis. Compared with the Feishu version, DingTalk's exclusive advantages: it can integrate OA approval efficiency analysis and attendance data.

Trigger Method

"商业分析"  "业务报告"  "顾问分析"  "夜间分析"  "业务建议"  "全面分析"

8 Advisor Roles

Role	Focus Area	Data Source	DingTalk Exclusive Enhancement
Financial Advisor	Revenue trends, investment performance, expenses	Investment Portfolio (aitable)	—
Marketing Advisor	Content performance, engagement rate, topic popularity	Social Tracking (aitable)	—
Growth Advisor	Follower growth, channel acquisition efficiency	Social Tracking (aitable)	—
Operations Advisor	Task completion rate, approval efficiency	To-dos + Calendar	OA approval efficiency
Content Strategy Advisor	Topic selection quality, publishing cadence	Creative Library + Social Tracking	—
Relationship/BD Advisor	Network activity, collaboration opportunities	CRM Contacts	—
Competition/Industry Advisor	Industry trends, competitor activity	Knowledge Base + WebSearch	—
Comprehensive Risk Advisor	Potential risks, weak links	All data	Attendance anomaly risk

Data Initialization

On first use, automatically create aitable data tables:

# 创建 Base
dws aitable base create --name "商业顾问团" --yes -f json
# 返回 baseId，记为 <ba-base-id>

# 创建 2 张表
dws aitable table create --base-id <ba-base-id> --name "分析记录" --yes -f json
# 返回 tableId，记为 <records-table-id>

dws aitable table create --base-id <ba-base-id> --name "用户反馈" --yes -f json
# 返回 tableId，记为 <feedback-table-id>

# 添加业务字段（表创建后自动带一个"标题"primaryDoc字段，需额外添加业务字段）
dws aitable field create --base-id <ba-base-id> --table-id <records-table-id> --fields '[{"fieldName":"分析日期","type":"date"},{"fieldName":"顾问角色","type":"text"},{"fieldName":"分析摘要","type":"text"},{"fieldName":"优先级","type":"text"},{"fieldName":"状态","type":"text"},{"fieldName":"详细分析","type":"text"},{"fieldName":"建议行动","type":"text"}]' --yes -f json

dws aitable field create --base-id <ba-base-id> --table-id <feedback-table-id> --fields '[{"fieldName":"分析日期","type":"date"},{"fieldName":"展开编号","type":"number"},{"fieldName":"用户意见","type":"text"}]' --yes -f json

Table 1: Analysis Records (advisor_records)

Field Name	Type	Description
Analysis Date	Date	Analysis execution date
Advisor Role	Text	One of 8 roles
Analysis Summary	Text	Key findings
Priority	Text	High/Medium/Low
Status	Text	Pending view/Expanded/Processed
Detailed Analysis	Text	Full analysis content after expansion
Recommended Actions	Text	Recommended action steps

Table 2: User Feedback (user_feedback)

Field Name	Type	Description
Analysis Date	Date	Associated analysis date
Expanded Number	Number	The suggestion number selected by the user
User Feedback	Text	Adopt/Ignore/Supplement

Working Modes

Mode 1: Full Analysis (8 advisors in parallel)

Trigger: "Help me analyze the recent business situation" / "Run the business advisory team" / scheduled automatic execution

Workflow:

Step 1: 并行拉取所有可用数据
├── dws todo task list → 待办任务数据
├── dws calendar event list → 日程数据
├── dws oa approval list-pending → OA 审批待办（钉钉独有）
├── dws oa approval list-initiated → 我发起的审批（钉钉独有）⚠️ 当前不可用，跳过
├── dws attendance record get → 考勤数据（钉钉独有）
├── dws aitable record query → CRM 联系人（如已初始化）
├── dws aitable record query → 社交追踪快照（如已初始化）
├── dws aitable record query → 创意库（如已初始化）
└── dws aitable record query → 投资数据（如已初始化）

Step 2: 8 个顾问角色并行分析（使用 Agent Tool）
├── 财务顾问 → 分析投资组合表现
├── 营销顾问 → 分析内容互动数据
├── 增长顾问 → 分析增长趋势
├── 运营顾问 → 分析任务完成率 + OA 审批效率（钉钉独有）
├── 内容策略顾问 → 分析创意和内容质量
├── 关系/BD 顾问 → 分析联系人动态
├── 竞争/行业顾问 → WebSearch 行业趋势
└── 综合风险顾问 → 综合评估所有风险（含考勤异常，钉钉独有）

Step 3: AI 汇总去重
├── 合并相似建议
├── 按优先级排序（高→中→低）
├── 编号（1, 2, 3...）
└── 标注来源顾问

Step 4: 生成编号清单

Step 5: 保存分析记录
└── dws aitable record create → 批量保存各顾问分析

Step 6: 发送清单
├── dws chat message send-by-bot → 发送到群聊
└── 或 dws ding message send → DING 紧急发送

Output format:

## 今日商业顾问团报告（2026-04-06）

### 总览
今日共 N 条建议，来自 M 位顾问。
数据来源：CRM(有)、社交追踪(有)、创意库(有)、投资(无)、OA审批(有)、考勤(有)

### 建议清单
1. 【关系】张三（XX科技）已 45 天未联系，上次讨论了合作意向，建议主动跟进。[来源: 关系/BD顾问]
2. 【财务】投资组合本周下跌 5%，科技股拖累，建议检查持仓比例。[来源: 财务顾问]
3. 【内容】本周内容发布频率下降 40%，建议安排时间完成待执行选题。[来源: 内容策略顾问]
4. 【运营】7 个任务已过期未完成，其中 3 个超过 14 天。[来源: 运营顾问]
5. 【审批】有 3 个审批即将到期，其中 1 个超 3 天未处理。[来源: 运营顾问]（钉钉独有）
6. 【考勤】本周有 2 天迟到，注意调整作息。[来源: 综合风险顾问]（钉钉独有）
7. 【风险】竞品 YY 产品本周发布新功能，与产品路线图重叠。[来源: 竞争顾问]
8. 【增长】Instagram 增长势头良好，建议加大投入。[来源: 增长顾问]

回复编号查看详细分析。

Mode 2: Expand detailed analysis

Trigger: "Tell me more about item 3" / "Expand item N"

Workflow:

Step 1: 查询对应顾问的完整分析记录
└── dws aitable record query --base-id <ba-base-id> --table-id <records-table-id> --yes -f json

Step 2: AI 生成展开分析
├── 原始建议
├── 来源顾问
├── 详细分析（数据支撑）
├── 建议行动步骤（3-5 步）
└── 预期效果

Step 3: 记录用户反馈
└── dws aitable record create --base-id <ba-base-id> --table-id <feedback-table-id> --records '[...]' --yes -f json

Expansion format:

## 第 3 条展开分析

### 原始建议
本周内容发布频率下降 40%，建议安排时间完成待执行选题。

### 来源顾问：内容策略顾问

### 详细分析
- 上周发布 5 条内容，本周仅 3 条
- 创意库中有 4 个"方案已完成"状态但未执行
- YouTube 表现最好的内容类型是教程类（互动率 4.2%）
- Instagram 最受欢迎的是图文干货（互动率 5.8%）

### 建议行动步骤
1. 优先执行创意库中评分最高的"AI编程工具对比"选题
2. 本周至少发布 2 条教程类内容
3. 安排 30 分钟完成"飞书自动化"选题的资料整理

### 预期效果
内容发布频率恢复到 5 条/周，预计互动率提升 15%。

Mode 3: Single advisor analysis

Trigger: "Just look at the financial advisor's analysis" / "What suggestions does the operations advisor have"

Workflow:

Step 1: 拉取指定顾问所需数据

Step 2: 执行单顾问分析

Step 3: 输出分析报告

Mode 4: Historical analysis review

Trigger: "What was in last week's advisor analysis" / "Review the previous analysis"

Workflow:

Step 1: 查询历史分析记录
└── dws aitable record query --base-id <ba-base-id> --table-id <records-table-id> --limit 100 --yes -f json

Step 2: AI 按日期分组，格式化输出

Step 3: 标注已处理/未处理状态

dws CLI command reference

Create business advisory team

dws aitable base create --name "商业顾问团" --yes -f json
dws aitable table create --base-id <ba-base-id> --name "分析记录" --yes -f json
dws aitable table create --base-id <ba-base-id> --name "用户反馈" --yes -f json

Pull business data

# 待办任务
dws todo task list --yes -f json

# 日程
dws calendar event list --yes -f json

# OA 审批（钉钉独有）
dws oa approval list-pending --yes -f json
# ⚠️ list-initiated 当前不可用（服务端参数错误）
# dws oa approval list-initiated --yes -f json

# 考勤（钉钉独有）
dws attendance record get --user "manager6950" --date "2026-04-06" --yes -f json
# ⚠️ attendance summary 当前不可用（服务端统计类型错误 C0002），可用 attendance record get 替代
# dws attendance summary --user "manager6950" --date "2026-04-06 00:00:00" --yes -f json

# 联动数据（需对应 Skill 已初始化）
dws aitable record query --base-id <crm-base-id> --table-id <contacts-table-id> --limit 100 --yes -f json
dws aitable record query --base-id <social-base-id> --table-id <snapshots-table-id> --limit 100 --yes -f json
dws aitable record query --base-id <cp-base-id> --table-id <ideas-table-id> --limit 100 --yes -f json
dws aitable record query --base-id <sa-base-id> --table-id <analysis-table-id> --limit 100 --yes -f json

Save analysis records

dws aitable record create --base-id <ba-base-id> --table-id <records-table-id> --records '[{"cells":{"标题":"运营顾问-任务过期分析","分析日期":"2026-04-06","顾问角色":"运营顾问","分析摘要":"7个任务过期未完成","优先级":"高","状态":"待查看"}}]' --yes -f json

Save user feedback

dws aitable record create --base-id <ba-base-id> --table-id <feedback-table-id> --records '[{"cells":{"标题":"用户反馈-第3条","分析日期":"2026-04-06","展开编号":3,"用户意见":"已采纳，本周安排"}}]' --yes -f json

Send report to group chat

dws chat message send-by-bot --robot-code "bot_code" --group "openConversationId" --title "商业顾问团报告" --text "今日共 8 条建议..." --yes

DING urgent send (high-priority suggestions)

dws ding message send --robot-code "bot_code" --users "manager6950" --content "【紧急】3个审批即将到期，请及时处理" --type "1" --yes

Scheduled tasks

Task	Cron	Description
Nightly analysis	`0 22 * * 1-5`	Automatically runs and pushes at 22:00 on workdays

CronCreate: cron="0 22 * * 1-5", prompt="运行商业顾问团分析，生成今日报告并发送到指定群聊。如有高优先级建议用DING发送", recurring=true, durable=true

Linked Skills (data dependencies)

Linked Skill	Provided data	Analysis dimension
personal-crm	Contact heat, interaction records	Relationship/BD advisor
social-tracker	Follower growth, content performance	Marketing/growth advisor
content-pipeline	Idea library, content quality	Content strategy advisor
stock-analyzer	Investment portfolio, analysis records	Financial advisor
meeting-todo	Meeting to-do completion rate	Operations advisor
health-diary	Health data	Comprehensive risk advisor

DingTalk-exclusive enhancements

OA approval efficiency analysis

Note: dws oa approval list-initiated The current server error makes it unavailable, so the following analysis cannot be executed for now. You can use dws oa approval list-pending(normally available) to obtain pending approval data for partial analysis.

运营顾问额外分析（当前 list-initiated 不可用，部分指标暂缺）：
- 本周新增审批 N 件，处理 M 件，处理率 X%  ← 依赖 list-initiated，暂不可用
- 平均处理时间 Y 小时  ← 依赖 list-initiated，暂不可用
- 超过 3 天未处理的审批 N 件  ← 可通过 list-pending 获取
- 被驳回的审批 N 件，主要原因：...  ← 依赖 list-initiated，暂不可用

Attendance risk analysis

综合风险顾问额外分析：
- 本周迟到 N 次，平均迟到 M 分钟
- 本月早退 N 次
- 工作时长趋势（是否过劳）
- 建议：注意调整作息，避免影响工作效率

Notes

All dws commands must include the --yes flag
The 8 advisors execute in parallel using Agent Tool, with a total duration of about 2-5 minutes
Linked data depends on other Skills already being initialized (aitable tables created); data from uninitialized Skills is marked as "none"
Summary deduplication is the core step: different advisors may propose similar suggestions, which need to be merged
Priority determination: those affecting revenue/cooperation are high, those affecting efficiency are medium, and optimization suggestions are low
DING is only used for high-priority suggestions, and group messages are used for regular reports
DingTalk's OA approval and attendance are exclusive data sources that other platforms cannot obtain
dws oa approval list-initiated and dws attendance summary currently have server errors; the business advisory team's operations advisor and comprehensive risk advisor need to skip these two data sources during analysis
dws attendance record get and dws attendance rules are normally available and can be used as alternative data sources

Known issues

dws oa approval list-initiated The server reports "parameter error": even when providing the --start/--end parameter, it still reports an error. Currently it is recommended to use dws oa approval list-pending(Normally available) alternative, or wait for the server-side fix.
dws attendance summary The server reports "statistics type error" (C0002): unavailable. For attendance data, switch to dws attendance record get, which allows querying day by day by date to obtain raw clock-in records.

About this skill

Trigger Method

8 Advisor Roles

Data Initialization

Working Modes

Mode 1: Full Analysis (8 advisors in parallel)

Mode 2: Expand detailed analysis

Mode 3: Single advisor analysis

Mode 4: Historical analysis review

dws CLI command reference

Create business advisory team

Pull business data

Save analysis records

Save user feedback

Send report to group chat

DING urgent send (high-priority suggestions)

Scheduled tasks

Linked Skills (data dependencies)

DingTalk-exclusive enhancements

OA approval efficiency analysis

Attendance risk analysis

Notes

Known issues