AgentHack submission type
Enterprise Agents
Name
Gaurav Choudhary
Team name
Ninja
Team members
priyanka.o.jain@accenture.com, mamta.sawle@accenture.com, manish.r.kuma@accenture.com, g.a.choudhary@accenture.com
How many agents do you use
Multiple agents
Industry category in which use case would best fit in (Select up to 2 industries)
Operations
Other sector
Complexity level
Advanced
Summary (abstract)
The current PR (Purchase Requisition) to PO (Purchase Order) process in many procurement environments is fragmented and heavily reliant on human judgment, especially in the supplier selection and verification stages. When multiple quotations are received, procurement staff must manually investigate and evaluate preferred suppliers based on various parameters such as past performance, delivery timelines, cost competitiveness, and quality metrics.
This decision-making process is time-consuming and requires experienced personnel to ensure accuracy and compliance. Additionally, once a PO is created, PO acknowledgment emails are manually sent to suppliers, followed by email tracking and reminders in case of no response. These tasks create bottlenecks, increase lead time, and introduce a risk of human error or communication gaps.
Despite some rule-based automations possibly existing for document handling or form filling, the judgment-intensive and communication-driven tasks remain manual, limiting end-to-end automation potential. The lack of intelligent decision-making and proactive follow-up are key pain points that inhibit full process transformation.
Detailed problem statement
Automation of PR to PO process is module specific and fragmented due to judgmental steps involved.
Agents are required to investigate for the preferred supplier which is a time-consuming task and has dependency on experienced & skilled staff.
In case multiple quotes are received along with the PR approval request, complex investigation is required for identifying recommended bidder with all required parameters.
The process takes ~~ 45 min manually for each supplier verification & recommendation in the overall PR to PO cycle.
Additionally, once a PO is created, PO acknowledgment emails are manually sent to suppliers, followed by email tracking and reminders in case of no response. These tasks create bottlenecks, increase lead time, and introduce a risk of human error or communication gaps.
Detailed solution
To address the above challenges, we propose an agentic automation solution using UiPath Autopilot and Agents to create an intelligent, end-to-end automated workflow for the PR to PO cycle.RPA to fetch data from sources and automate std steps. Below are few Agents with which we are developing our solution
Quote Validator Agent :Validates if received quotations match PR details — ensuring correctness in quantities, items, and pricing before further processing.
Preferred & Non-preferred Supplier Agent :Identifies preferred suppliers for the requested items using historical data and predefined lists, supporting compliance and consistency.
Quote Recommendation Agent: Analyzes validated quotes on cost, quantity, and compliance to suggest the most suitable options.
Supplier Recommender Agent: Combines preferred, non-preferred, and historical supplier data to recommend the best supplier when no preferred option fits.
Demo Video
Expected impact of this automation
Implementing this agentic automation solution is expected to generate significant business impact, including:
• Time Savings: Reduction of manual effort by over 60-70%, cutting down supplier recommendation time from ~30 minutes to <10 minutes per instance.
• Resource Optimization: Freeing up skilled procurement staff to focus on strategic tasks rather than routine investigation and follow-up.
• Improved Accuracy: Decision-making based on standardized logic and data-driven insights reduces errors and inconsistencies.
• Enhanced Responsiveness: Automated supplier follow-ups improve communication turnaround and PO cycle timelines.
• Scalability: The modular design allows deployment across different departments, business units, or procurement categories without extensive rework.
• Intelligent Decision Support: Combines automation with human validation, enhancing trust in AI-driven recommendations.
Estimated Savings:-
70+ Clients
3000+ FTEs
1000+ Addressable FTEs
40% Productivity
400+ FTE save assumed
UiPath products used (select up to 4 items)
UiPath Agent Builder
UiPath Apps
UiPath Integration Service
UiPath Maestro
Integration with external technologies
gpt-4o
Agentic solution architecture (file size up to 4 MB)
Sample inputs and outputs for solution execution
2596188.pdf
2750899_1.pdf
2750899_2.pdf
MailData.xlsx
MaterialMaster.xlsx
PO1111.doc
pohistory.xlsx
PRDumpNew.xlsx
sample_explain.pdf