Skip to content

Commit da5f6d4

Browse files
authored
Merge pull request #211754 from amsliu/access-reviews-update
public preview update
2 parents 725dfb1 + 58e9d2c commit da5f6d4

File tree

3 files changed

+23
-4
lines changed

3 files changed

+23
-4
lines changed

articles/active-directory/governance/create-access-review.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -139,7 +139,7 @@ If you are reviewing access to an application, then before creating the review,
139139

140140
1. In the **Enable review decision helpers** section choose whether you want your reviewer to receive recommendations during the review process:
141141
1. If you select **No sign-in within 30 days**, users who have signed in during the previous 30-day period are recommended for approval. Users who haven't signed in during the past 30 days are recommended for denial. This 30-day interval is irrespective of whether the sign-ins were interactive or not. The last sign-in date for the specified user will also display along with the recommendation.
142-
1. If you select User-to-Group Affiliation, reviewers will get the recommendation to Approve or Deny access for the users based on user’s average distance in the organization’s reporting-structure. Users who are very distant from all the other users within the group are considered to have "low affiliation" and will get a deny recommendation in the group access reviews.
142+
1. If you select **(Preview) User-to-Group Affiliation**, reviewers will get the recommendation to Approve or Deny access for the users based on user’s average distance in the organization’s reporting-structure. Users who are very distant from all the other users within the group are considered to have "low affiliation" and will get a deny recommendation in the group access reviews.
143143

144144
> [!NOTE]
145145
> If you create an access review based on applications, your recommendations are based on the 30-day interval period depending on when the user last signed in to the application rather than the tenant.
189 KB
Loading

articles/active-directory/governance/review-recommendations-access-reviews.md

Lines changed: 22 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -10,7 +10,7 @@ ms.workload: identity
1010
ms.tgt_pltfrm: na
1111
ms.topic: how-to
1212
ms.subservice: compliance
13-
ms.date: 8/5/2022
13+
ms.date: 10/25/2022
1414
ms.author: amsliu
1515
ms.reviewer: mwahl
1616
ms.collection: M365-identity-device-management
@@ -29,7 +29,26 @@ For more information, see [License requirements](access-reviews-overview.md#lice
2929
## Inactive user recommendations
3030
A user is considered 'inactive' if they have not signed into the tenant within the last 30 days. This behavior is adjusted for reviews of application assignments, which checks each user's last activity in the app as opposed to the entire tenant. When inactive user recommendations are enabled for an access review, the last sign-in date for each user will be evaluated once the review starts, and any user that has not signed-in within 30 days will be given a recommended action of Deny. Additionally, when these decision helpers are enabled, reviewers will be able to see the last sign-in date for all users being reviewed. This sign-in date (as well as the resulting recommendation) is determined when the review begins and will not get updated while the review is in-progress.
3131

32+
## User-to-Group Affiliation (preview)
33+
Making the review experience easier and more accurate empowers IT admins and reviewers to make more informed decisions. This Machine Learning based recommendation opens the journey to automate access reviews, thereby enabling intelligent automation and reducing access rights attestation fatigue.
34+
35+
User-to-Group Affiliation in an organization’s chart is defined as two or more users who share similar characteristics in an organization's reporting structure.
36+
37+
This recommendation detects user affiliation with other users within the group, based on organization's reporting-structure similarity. The recommendation relies on a scoring mechanism which is calculated by computing the user’s average distance with the remaining users in the group. Users who are very distant from all the other group members based on their organization's chart, are considered to have "low affiliation" within the group.
38+
39+
If this decision helper is enabled by the creator of the access review, reviewers can receive User-to-Group Affiliation recommendations for group access reviews.
40+
41+
> [!NOTE]
42+
> This feature is only available for users in your directory. A user should have a manager attribute and should be a part of an organizational hierarchy for the User-to-group Affiliation to work.
43+
44+
The following image has an example of an organization's reporting structure in a cosmetics company:
45+
46+
![Screenshot that shows a fictitious hierarchial organization chart for a cosmetics company.](./media/review-recommendations-group-access-reviews/org-chart-example.png)
47+
48+
Based on the reporting structure in the example image, users who are statistically significant amount of distance away from other users within the group, would get a "Deny" recommendation by the system if the User-to-Group Affiliation recommendation was selected by the reviewer for group access reviews.
49+
50+
For example, Phil who works within the Personal care division is in a group with Debby, Irwin, and Emily who all work within the Cosmetics division. The group is called *Fresh Skin*. If an Access Review for the group Fresh Skin is performed, based on the reporting structure and distance away from the other group members, Phil would be considered to have low affiliation. The system will create a **Deny** recommendation in the group access review.
51+
3252
## Next Steps
3353
- [Create an access review](create-access-review.md)
34-
- [Review access to groups or applications](perform-access-review.md)
35-
54+
- [Review access to groups or applications](perform-access-review.md)

0 commit comments

Comments
 (0)