How SoftBank Uses RPA and Generative AI to Save 4,500 FTEs Annually

SoftBank Corp. is Japan’s largest telecommunications operator and a global technology investor with ambitions far beyond carrier services. Facing a bottleneck in scaling automation beyond early RPA deployments, SoftBank launched the ‘Digital Worker 4,000 Project’—a top-management-led initiative to automate cross-functional business processes using Automation Anywhere alongside Salesforce, Office365, and Google integrations. The effort redirected 4,500 FTE-equivalents of annual labor toward value-added work and generated 52,000 generative AI improvement proposals from employees in just ten days.

Impact

4,500

FTE-equivalents saved annually

85%

Reduction in recruitment hours

50%

Reduction in mobile service registration time

700 hours

AI call volume prediction savings

52,000

Employee generative AI proposals submitted

Challenge

SoftBank had deployed RPA successfully in isolated pockets since 2016, but could not scale automation across the enterprise—manual processes were consuming millions of hours annually and the existing task-based approach could not address complex, multi-system workflows.

Solution

SoftBank launched the ‘Digital Worker 4,000 Project’ using Automation Anywhere to automate cross-functional workflows across Salesforce, Office365, and Google Workspace, paired with a company-wide employee training program to drive bottom-up automation identification and adoption.

Tools & Technologies

What Leaders Say

We aimed to inspire a fresh sensibility through the experience of technology… we have been realizing a work style and workplace where employees and technology are integrated together.

Satoshi Kaminagayoshi, VP of Customer Success, SoftBank Corp.

Generative AI will continue to merge with existing technologies. I see this trend accelerating the realization of end-to-end process automation, moving away from automation based on task.

Takashi Kinoshita, Director, Product & Business Strategy, SoftBank Corp.
Get the full story.

Sign up to read complete case studies, access detailed metrics, and unlock all use cases.

Full Story

SoftBank Corp. serves tens of millions of subscribers across Japan and operates as a strategic technology investor globally. The company began its automation journey in 2016 with RPA, but scaling that initial work proved harder than expected. Isolated deployments, inconsistent training, and a task-by-task approach left the company short of its goal: transforming how employees experience their work.

The bottleneck was structural. Automation tools existed, but they weren’t embedded into the fabric of how teams operated. Manual, repetitive processes—from call volume prediction to mobile service registration—continued to consume capacity that SoftBank needed for higher-value activities. The company set an ambitious target: save 7.7 million hours per year while simultaneously improving both customer and employee experience.

SoftBank launched the ‘Digital Worker 4,000 Project,’ led directly by top management. The initiative used Automation Anywhere to orchestrate complex cross-functional processes spanning Salesforce, Office365, Google Workspace, and internal systems. Alongside deployment, SoftBank ran a company-wide upskilling program to give employees the skills to identify and implement their own automations—turning automation from an IT initiative into a workforce capability.

The results went beyond the original metrics. Recruiting evaluations now take 85% fewer hours. Mobile service registration is 50% faster. AI-powered call volume prediction saves 700 hours. And when SoftBank held an internal generative AI ideas contest, employees submitted 52,000 proposals in just ten days—a signal that the culture had fundamentally shifted toward technology-first problem solving.

SoftBank is now moving beyond task-level automation toward end-to-end process automation powered by generative AI and agentic systems. The company is testing AI-generated insights for strategic decision-making and intends to extend this infrastructure to customers—positioning itself not just as a telecom carrier, but as a digital transformation enabler for its enterprise base.

Similar Cases

C
Cisco
1,500+
engineering hours saved monthly

Cisco embedded OpenAI Codex into production engineering workflows, achieving 20% faster builds, 10-15x defect resolution throughput, and saving 1,500+ engineering hours per month.

TelecommunicationsOCOpenAI Codex
RU
RMIT University
60,000+
total staff hours returned (3 years)

RMIT University is one of Australia’s largest and most globally connected universities, serving hundreds of thousands of students across campuses in Melbourne, Vietnam, Spain, and partner institutions worldwide. Facing high volumes of international student applications and administrative workflows spread across multiple disconnected systems, RMIT’s automation team deployed 27 automation solutions using Automation Anywhere’s platform, including five AI-powered automations. Over three years, RMIT returned more than 60,000 staff hours to the institution—equivalent to 24 years of capacity—while processing more than 20,000 student requests with greater speed and accuracy.

EducationAAAutomation Anywhere
SB
Shinhan Bank
1.2M+
staff hours saved annually

Shinhan Bank is one of South Korea’s largest financial institutions, operating more than 800 branches and serving millions of customers. Facing a workforce burdened by repetitive compliance, loan review, and data-transfer tasks across siloed systems, the bank launched the ‘My Bot’ initiative—deploying personal digital assistants to all 14,000 employees using Automation Anywhere’s platform. The result is one of the world’s largest enterprise digital assistant deployments, saving over 1.2 million staff hours annually and automating more than 100,000 property loan reviews each year.

Financial ServicesAAAutomation Anywhere
TW
The Washington Post
100%
invoice tax accuracy coverage

The Washington Post is a technology-forward media company combining world-class journalism with digital innovation. Facing an unmanageable volume of vendor invoices with inconsistent tax formats, The Post deployed an AI Agent powered by proprietary large language models through Automation Anywhere’s Agentic Process Automation platform. The result: 100% of invoices are now validated for tax accuracy, tax overpayments have been eliminated entirely, and the Finance team captured $1 million in automation value within year one.

Media & PublishingAAAutomation Anywhere