You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: learn-pr/wwl/analyze-requirements-ai-powered-business-solutions/2-assess-use-agents-task-automation-data-analytics-decision-making.yml
Copy file name to clipboardExpand all lines: learn-pr/wwl/analyze-requirements-ai-powered-business-solutions/3-review-data-grounding-accuracy-relevance-timeliness-cleanliness-availability.yml
Copy file name to clipboardExpand all lines: learn-pr/wwl/analyze-requirements-ai-powered-business-solutions/4-organize-business-solution-data-available-other-ai-systems.yml
Copy file name to clipboardExpand all lines: learn-pr/wwl/analyze-requirements-ai-powered-business-solutions/includes/2-assess-use-agents-task-automation-data-analytics-decision-making.md
+2-2Lines changed: 2 additions & 2 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -94,6 +94,6 @@ These principles reinforce reliable, secure AI adoption at scale.
94
94
95
95
Use these links as the primary sources for this unit:
Copy file name to clipboardExpand all lines: learn-pr/wwl/analyze-requirements-ai-powered-business-solutions/includes/3-review-data-grounding-accuracy-relevance-timeliness-cleanliness-availability.md
+13-13Lines changed: 13 additions & 13 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -9,11 +9,11 @@ After completing this unit, learners will be able to:
9
9
10
10
## 1. Understanding Grounding in AI Agents
11
11
12
-
Grounding ensures that an AI agent responds using **trusted, domainspecific organizational data**, increasing accuracy and reducing hallucinations.
12
+
Grounding ensures that an AI agent responds using **trusted, domain-specific organizational data**, increasing accuracy and reducing inaccurate information.
13
13
14
14
Microsoft Copilot and Copilot Studio use **semantic indexing** to map enterprise content across Microsoft Graph into rich lexical and semantic representations. This enables more contextually precise retrieval.
15
15
16
-
AI systems must be connected to **approved, accesscontrolled data** so they produce trustworthy outcomes that respect organizational security boundaries.
16
+
AI systems must be connected to **approved, access-controlled data** so they produce trustworthy outcomes that respect organizational security boundaries.
17
17
18
18
To support advanced grounding, the **Copilot Retrieval API** retrieves relevant text passages from SharePoint, OneDrive, and connected sources, honoring user permissions.
19
19
@@ -89,7 +89,7 @@ The Retrieval API respects permissions and will not return content beyond a user
89
89
90
90
|**Dimension**|**Definition**|**Impact on Agent Performance**|
91
91
|---|---|---|
92
-
|**Accuracy**| Data is correct and verified | Reduces hallucinations and misinformation |
92
+
|**Accuracy**| Data is correct and verified | Reduces inaccurate information and misinformation |
93
93
|**Relevance**| Data aligns to the task/intent | Ensures responses match the intended scenario |
94
94
|**Timeliness**| Data is current and up to date | Keeps outputs aligned with latest policies or info |
95
95
|**Cleanliness**| Data is structured and free of noise | Improves retrieval precision |
@@ -101,24 +101,24 @@ The Retrieval API respects permissions and will not return content beyond a user
101
101
102
102
## 5. Best Practices for Reviewing Grounding Data
103
103
104
-
1.**Evaluate content quality before upload**: remove outdated or conflicting information.
104
+
-**Evaluate content quality before upload**: remove outdated or conflicting information.
105
105
106
-
2.**Store authoritative content in SharePoint or OneDrive** so it becomes part of the semantic index.
106
+
-**Store authoritative content in SharePoint or OneDrive** so it becomes part of the semantic index.
107
107
108
-
3.**Ensure consistent formatting** to improve data cleanliness and retrieval precision.
108
+
-**Ensure consistent formatting** to improve data cleanliness and retrieval precision.
109
109
110
-
4.**Review permissions regularly** so agents ground from valid data sources only.
110
+
-**Review permissions regularly** so agents ground from valid data sources only.
111
111
112
-
5.**Collaborate with domain SMEs** to validate accuracy and contextual fit.
112
+
-**Collaborate with domain SMEs** to validate accuracy and contextual fit.
113
113
114
114
**References**
115
115
116
-
_Add knowledge to ground an agent - Copilot Studio_<br>[Add knowledge to ground an agent](/training/modules/agents-copilot-studio-online-workshop/add-knowledge)
116
+
-[Add knowledge to ground an agent](/training/modules/agents-copilot-studio-online-workshop/add-knowledge)
117
117
118
-
_Ground AI using trusted data_<br>[Ground AI using trusted data](/training/modules/build-effective-generative-ai-solutions-organization/3-ground-ai-using-trusted-data)
118
+
-[Ground AI using trusted data](/training/modules/build-effective-generative-ai-solutions-organization/3-ground-ai-using-trusted-data)
119
119
120
-
_Semantic indexing for Microsoft 365 Copilot_<br>[Semantic indexing for Microsoft 365 Copilot](/microsoftsearch/semantic-index-for-copilot)
120
+
-[Semantic indexing for Microsoft 365 Copilot](/microsoftsearch/semantic-index-for-copilot)
_Manage Grounding With Bing in Microsoft Foundry and Azure_<br>[Manage Grounding With Bing](/azure/ai-foundry/agents/how-to/manage-grounding-with-bing)
124
+
-[Manage Grounding With Bing in Microsoft Foundry and Azure](/azure/ai-foundry/agents/how-to/manage-grounding-with-bing)
Copy file name to clipboardExpand all lines: learn-pr/wwl/analyze-requirements-ai-powered-business-solutions/includes/4-organize-business-solution-data-available-other-ai-systems.md
+11-11Lines changed: 11 additions & 11 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -2,14 +2,14 @@
2
2
3
3
By the end of this unit, learners will be able to:
4
4
5
-
- Explain why wellorganized business solution data is essential for AI readiness.
5
+
- Explain why well-organized business solution data is essential for AI readiness.
6
6
- Identify architectural components that enable AI agents and AI systems to consume organizational data.
7
7
- Apply best practices for structuring, storing, indexing, and exposing data so AI systems can use it reliably.
8
8
- Understand how the Azure platform, Microsoft databases, and data architecture patterns support enterprise AI scenarios.
9
9
10
10
## 1. Introduction
11
11
12
-
AI systems—whether Copilot, autonomous agents, or custom-built AI applications—require **highquality, structured, and accessible data**. Poorly organized data leads to weak grounding, hallucinations, data quality issues, and unreliable decisionmaking.
12
+
AI systems—whether Copilot, autonomous agents, or custom-built AI applications—require **high-quality, structured, and accessible data**. Poorly organized data leads to weak grounding, inaccurate information, data quality issues, and unreliable decision making.
13
13
14
14
Organizing business solution data is not only a technical requirement but a business imperative. When data is structured correctly, it becomes available for:
15
15
@@ -19,27 +19,27 @@ Organizing business solution data is not only a technical requirement but a busi
19
19
20
20
- Custom AI apps built with Azure AI
21
21
22
-
-Retrievalaugmented generation (RAG) pipelines
22
+
-Retrieval augmented generation (RAG) pipelines
23
23
24
24
- Analytics and automation solutions
25
25
26
-
Retrieval-Augmented Generation (REG) is an architecture that separates prototypes from trustworthy systems. A RAG pipeline is the system that performs all the steps required to make RAG work in a production environment, handling the data ingestion, streaming, cleaning, chunking, embedding, indexing, retrieval, prompt assembly, orchestration, and monitoring that allow an LLM to use retrieved context when generating an answer There are several advantages of leveraging RAG pipelines:
26
+
RetrievalAugmented Generation (REG) is an architecture that separates prototypes from trustworthy systems. A RAG pipeline is the system that performs all the steps required to make RAG work in a production environment, handling the data ingestion, streaming, cleaning, chunking, embedding, indexing, retrieval, prompt assembly, orchestration, and monitoring that allow an LLM to use retrieved context when generating an answer There are several advantages of leveraging RAG pipelines:
27
27
28
28
- Empowering LLM solutions with real-time data access
29
29
30
30
- Preserving data privacy
31
31
32
-
- Mitigating LLM hallucinations
32
+
- Mitigating LLM inaccurate information
33
33
34
34
This unit explains how to organize your business data to become **usable, discoverable, secure, and optimized for AI consumption across the organization**.
35
35
36
-
## 2. Key Concepts for Organizing AIReady Data
36
+
## 2. Key Concepts for Organizing AI-Ready Data
37
37
38
38
### 2.1 Drive Transformation with Azure Data & AI Platforms
39
39
40
40
_(From: Leverage AI tools - Drive Transformation on Azure)_
41
41
42
-
Azure provides the foundational components necessary to **centralize, transform, and govern data** before AI systems consume it.Key concepts include:
42
+
Azure provides the foundational components necessary to **centralize, transform, and govern data** before AI systems consume it.Key concepts include:
43
43
44
44
-**Unified data estate**: Consolidate data from apps, logs, CRM, ERP, operations, and documents.
45
45
@@ -68,19 +68,19 @@ Microsoft's Cloud Adoption Framework defines how to design a **data architecture
68
68
69
69
-**Semantic indexing**: Converts enterprise content into semantic representations for grounding.
70
70
71
-
-**Data governance layer**: Rolebased access, sensitivity labels, Microsoft Purview.
71
+
-**Data governance layer**: Role-based access, sensitivity labels, Microsoft Purview.
72
72
73
73
-**APIs and connectors**: Ensure AI agents can access structured and unstructured data.
0 commit comments