AgentHack submission type
Enterprise Agents
Name
Leonardo Delgado Rios
How many agents do you use
Multiple agents
Industry category in which use case would best fit in (Select up to 2 industries)
Finance
Operations
Compliance
Complexity level
Advanced
Summary (abstract)
This submission presents an intelligent automation workflow for the Procure-to-Pay (P2P) process, built on UiPath Maestro and Agent Builder. It combines a Request Validator Agent and a Stock & Fulfillment Planning Agent to autonomously evaluate procurement requests, enforce policy compliance and generate actionable fulfillment plans based on real time stock and shipment conditions. The system intelligently determines whether requests can be fulfilled internally or require supplier onboarding (start licitation, onboarding process, etc) triggering smart notifications to administrators and internal teams, without duplicating control logic. Designed for reliability, modularity, and real world enterprise scalability, this agentic automation offers a pragmatic yet innovative leap in orchestrating P2P decisions across departments.
Detailed problem statement
In large organizations, the Procure-to-Pay (P2P) process is often fragmented, manual and error prone leading to delays, policy violations, untracked requests and/or procurement bottlenecks. One of the key challenges lies in validating procurement requests against complex internal policies, real time stock availability and variable shipment conditions. Most systems lack the intelligence to determine whether a request should be fulfilled from existing inventory or escalated for supplier onboarding, nor do they handle ambiguity, edge cases or exceptions without manual intervention. On the other hand, some other systems offers pricy and general solutions that do not handle internal and granular specifications. This automation addresses the inefficiency and inconsistency of traditional procurement flows by embedding intelligent agents that autonomously validate requests, interpret procurement and shipment policies, as well as generate structured fulfillment plans. It not only identifies whether a request is actionable or requires administrative escalation, but also ensures downstream systems like Order-to-Cash are notified appropriately. By doing so, it reduces processing time, eliminates manual follow-ups, increases policy compliance and creates a scalable foundation for intelligent decision-making in enterprise procurement operations.
Detailed solution
To solve the complexity and inconsistency of the Procure-to-Pay (P2P) process, we implemented an agent-based automation architecture orchestrated through UiPath Maestro. The solution is powered by two purpose-built agents Request Validator Agent and Stock & Fulfillment Planning Agent, developed using Agent Builder and governed through AI Trust Layer Policies for full transparency and control. The Request Validator Agent interprets and validates incoming procurement requests against dynamic internal policies, which are defined and managed via Indexes provided by multiple Storage Buckets and business-rule alignment. It determines if the request is valid, complete and the user is authorized to initiate the request, also applies edge-case logic to escalate ambiguous or non-compliant scenarios to an administrator, as well as the user that originate the request. If the request is denied by the system or is canceled by the user it sends a notification email and ends the process. If the request is approved, control is passed to the Stock & Fulfillment Planning Agent, which reviews real-time inventory data and shipment policies. This agent intelligently determines whether the request can be fulfilled from current stock, what the delivery timeline would be under various shipment scenarios (e.g., local vs. imported goods, weekday working hours) and whether supplier onboarding is needed. While it does not initiate onboarding flows or Order-to-Cash processes, it triggers structured notifications to procurement administrators when external sourcing is required. Similarly, if the order is valid and fulfillable, the agent sends a notification to the Order-to-Cash team to continue the downstream financial process. The entire agent interaction is orchestrated using UiPath Maestro, ensuring seamless coordination, task routing and context preservation across agents. Components were developed using Studio Web to maintain cloud-native simplicity and rapid deployment. This layered and modular architecture enables autonomous decision-making, reduces human error and establishes a policy-driven foundation for scaling enterprise procurement automation.
Demo Video
Expected impact of this automation
This automation is designed to significantly improve the speed, accuracy and compliance of the Procure-to-Pay (P2P) process by replacing manual validation, stock review and escalation routines with intelligent, policy-aware agents. By autonomously validating requests and generating fulfillment plans based on real-time inventory and shipment policies, the solution reduces manual review time by up to 80%, especially for recurring or low-risk requests. It minimizes policy violations and unauthorized purchases through consistent enforcement of rules defined in the currenty Policies for Request Approval, leading to an estimated 30-50% improvement in compliance across centers. The orchestration of agents through Maestro eliminates delays between validation and fulfillment steps, enabling end-to-end decision cycles to occur in minutes instead of hours or days, as well as error handling since it could be taken from the error instead of from the beginning. Notifications to administrators and the Order-to-Cash team are automated, eliminating the need for redundant follow-ups or email chains. From a financial perspective, faster cycle times and fewer errors reduce operational costs and improve cash flow visibility. Overall, this automation not only enhances operational efficiency and governance but also frees up procurement and administrative staff to focus on higher-value strategic tasks. Finally, this automation sets the foundation in order to expand and continue automating the process end-to-end, leveraging the integration of Document Understanding, or even IXP, for document processing that could potentially need to be read across the following phases.
UiPath products used (select up to 4 items)
UiPath Action Center
UiPath Apps
UiPath Orchestrator
UiPath Studio Web
UiPath Agent Builder
UiPath Maestro
UiPath Integration Service
UiPath Automation Cloud™
UiPath Automation Ops
Integration with external technologies
GenAI Activities, Gmail, Drive
Agentic solution architecture (file size up to 4 MB)
Agentic P2P Flow.uip (8.4 KB)
Sample inputs and outputs for solution execution
Inputs : {in_userEmail: hopkinsmichael@owens-daniel.com
in_requestMessage: Requesting: 10 books and 5 computers.}
Output: {
“supplier_found”: true,
“quantity_ok”: false,
“supplier_matches”: [
“Gamma Supply”,
“Alpha Supplies”
],
“fulfillment_plan”: {
“plan_type”: “Partial”,
“step_1”: {
“description”: “Deliver 10 books from Gamma Supply and Alpha Supplies.”,
“delivery_days”: 3,
“assigned_to”: “Derek Zuniga”
},
“step_2”: {
“description”: “Acquire 5 computers through procurement.”,
“assigned_to”: “Kimberly Adkins”,
“estimated_procurement_days”: 7
}
},
“summary”: {
“total_items”: 15,
“supplier_count”: 2,
“delivery_window”: “3 days for books, 7 days for computers procurement.”
}
}