Invoice processing · Production

Uber and Accenture build a global RPA program on UiPath, saving an estimated $10 million per year

The problem

Uber's rapid global growth created rising operational costs and staffing pressure, with geographically fragmented processes—such as different invoice procedures across the US, UK, and China—making it hard to maintain regulatory compliance and give senior management a unified view of operations.

Workflow diagram · grounded in source
1
Stakeholder automation intake
trigger
“business stakeholders in constant communication with developers, making changes on the fly”
2
Priority pipeline refresh
routing
“the RPA team refreshes the priority pipeline every quarter or every six months. That way, automation requests in the queue remain fresh and don't become obsolete before they are begun.”
3
Bot development via incubation pod
ai_action
“Using UiPath as the core automation platform, the Accenture team continues to update automations. They're continually adding new capabilities such as chatbots, optical character recognition (OCR), and machine learning.”
4
Bot identity and enterprise access
integration
“create bot ID processes that could accommodate unattended single sign-on to the enterprise”
5
Security and governance controls
validation
“Uber implemented rules to ensure secure access, such as logging in from recognized production machines, creating a secret log to store bot credentials, and regularly rotating passwords”
6
Diagnostic dashboards and ROI metrics
output
“dashboards for health monitoring and triage to identify problems before they become trouble tickets. These empowered the Operations and Support team to identify and correct problems as they arose.”
7
Stakeholder feedback incorporation
feedback_loop
“All the principles have been toward extensive communications and making sure we are getting feedback and incorporating that feedback from all stakeholders”
Reported outcome

After three years, Uber has more than 100 automations in production saving an estimated $10 million per year.
Uber Freight's invoice automation rate grew from less than 20% to over 70% of portal invoices monthly, and Uber avoided the potential loss of its London freight license, which would have cost 3% to 4% of total revenue.

Reported metrics
Automations in productionmore than 100
Estimated annual cost savings$10 million per year
Portal invoice automation rate (current)over 70%
Portal invoice automation rate (prior)less than 20%
Show all 11 reported metrics
automations in productionmore than 100
estimated annual cost savings$10 million per year
portal invoice automation rate (current)over 70%
portal invoice automation rate (prior)less than 20%
customers automated across portalsover 80 customers across seven main portals
Uber Freight customers and shippers handlednearly 100 customers and shippers
revenue at risk (London freight license)3% to 4% of the company's total revenue
time to build London safety tracking botsix weeks
invoice processing errorsdecreasing errors
customer satisfactionraises customer satisfaction
employee satisfactionimproved employee satisfaction
Reported stack
UiPathOCRchatbotsmachine learning
Source
https://www.uipath.com/resources/automation-case-studies/uber-maintains-global-infrastructure-built-on-rpa
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

After three years, Uber has more than 100 automations in production saving an estimated $10 million per year.

What tools did this team use?

UiPath, OCR, chatbots, machine learning.

What results were reported?

Automations in production: more than 100; Estimated annual cost savings: $10 million per year; Portal invoice automation rate (current): over 70%; Portal invoice automation rate (prior): less than 20% (source-reported, not independently verified).

How is this invoice processing AI workflow structured?

Stakeholder automation intake → Priority pipeline refresh → Bot development via incubation pod → Bot identity and enterprise access → Security and governance controls → Diagnostic dashboards and ROI metrics → Stakeholder feedback incorporation.

Related invoice processing cases