Identify Similar Records uses AI search to proactively surface related or duplicate demand records, helping Demand Managers maintain a cleaner, more accurate backlog.

 

With a simple button click on any demand record, the system analyzes name, description, and business case content against five years of historical data, identifying demands with 85% or higher semantic similarity. Results appear instantly in a global alert and related list, enabling Demand Managers to review potential duplicates and consolidate work before it impacts downstream processes.

 

By reducing unintentional duplication, this capability improves backlog quality, streamlines the approval process, and ensures resources focus on unique business initiatives. Demand Managers save time identifying redundancies; Portfolio Managers gain clearer visibility for prioritization and approvers make faster, more informed decisions.

 

Why It Matters:

 

Demand Managers often lack visibility into existing related demands when reviewing new submissions. This leads to unintentional duplication of work, which clutters the backlog, reduces data quality, and makes it harder for Portfolio Managers and approvers to prioritize effectively. Without an easy way to identify similar requests, duplicate demands slip through the review process, wasting resources, and creating confusion downstream.

 

What ** Makes *** It Special:*

 

• Cleaner demand backlog: Only unique requests in the backlog, eliminating duplicate clutter.

• Reduced duplication of effort: Teams avoid working on the same initiative twice under different demand records.

• Faster demand review and approval cycles: Approvers make quicker decisions with better visibility into existing work.

• Better portfolio visibility for decision-makers: Accurate view of planned work enables better prioritization and resource allocation.

This feature is turned on by default for customers on Zurich Patch 4 and above. If you have Now Assist with an appropriate role (Demand_user), you're ready to go.

 

How It Works** :**

 

1. Navigate to Demand Workbench and open a demand record.
2. Fill in the Name, Description and Business case fields to enhance the quality of the similarity check.
3. Click  Save  in case you’ve made updates.
4. Click " Identify similar demands" button from the demand header.
5. Review the global alert showing similar demands found (if any).
6. Scroll down to the  Similar Demands  related list.
7. Review each similar demand to determine if it's a duplicate. 
8. Consolidate or defer duplicates as needed.

If no similar records found:  A "No similar demands identified by Now Assist." message displays.

 

 

 

 

Give it a try and let us know how it goes. Have feedback or questions? Drop them in the comments below!

---

Frequently Asked Questions:

 

1. How does the AI determine if demands are similar?

Ans:  The AI uses semantic similarity analysis across the name, description, and business case fields in the demand backlog. Any demand with 85% or higher similarity from the past 5 years is flagged as similar.

 

2. What role is required to use Identify Similar Demands?

Ans:  Role required: Demand_user

 

3. How is the similarity check triggered?

Ans:  This capability has a manual trigger. Click the "Identify similar demands" button on the demand record header, to search for similar demands. This gives you full control over when to perform the search.

 

4. Where do similar demands appear after the search? 

Ans:  Similar demands appear in two places: 1. A global alert at the top of the demand record header listing the similar demands found, and 2. A "Similar Demands" related list when you scroll down on the demand record.

 

5. What happens if no similar demands are found? 

Ans:  A clear “No similar records found” message is displayed, confirming that no high-confidence matches were identified.

 

6. Are similar matches automatically merged or closed? 

Ans:  No. The system only surfaces potential matches. Demand Managers review the AI-generated list and decide whether to consolidate, relate, or delete duplicates.