Senior Writer

ADP’s cloud transformation pays dividends

16 Feb 20247 mins
Cloud ComputingDigital Transformation

The payroll giant has leveraged its early bets on cloud-native development to transform into an HCM provider poised to make the most of generative AI.

Vipul Nagrath stylized
Credit: Vipul Nagrath / ADP

For most organizations, a shift to the cloud brings scalability, access to innovative tools, and the possibility of cost savings. For payroll services company ADP, it has paved the way to becoming a SaaS provider capable of taking on big names in enterprise software.

An early partner of Amazon, the Roseland, N.J.-based company migrated to AWS in lockstep with Amazon’s earliest code releases a decade ago, even as the $18 billion payroll giant continued to build a hybrid cloud infrastructure that also incorporates Azure, Google Cloud Platform, and Cisco Cloud workloads.

Vipul Nagrath, who was global CIO of ADP during the company’s early embrace of the cloud, spearheaded ADP’s companywide digital transformation, which started in earnest five years ago fueled by the company’s hybrid cloud.

ADP’s aggressive digital transformation has not only cut costs and enabled more innovation but, most importantly, it has facilitated the payroll administrator’s evolution into a human capital management (HCM) service provider, which provides services to its customers from “hire to retire,” Nagrath says. These HCM services include applicant tracking, compensation, talent, and learning management, as well as insurance and retirement services.

“ADP remains the 500-pound gorilla in payroll and, with its rich data, can literally tell you what’s really moving the economy,” said Pete A. Tiliakos, principal analyst and managing partner of 3Sixty Insights, speaking to from an ADP conference in London in early February. “But about 10 years ago when they shifted towards being more of a technology services provider, they became a HCM service provider. That big cloud investment allowed them to begin competing against companies like Workday, SAP, and Oracle.”

The cloud-native advantage

ADP’s aggressive, early digital transformation has paid off nicely: Its expanded HCM portfolio is served to more than 1 million customers globally, up from 800,000 several years ago, with revenues at $18 billion in fiscal year 2023, up from $13 billion five years prior.

The 75-year-old company deeply prizes its close partnership with AWS but will continue to run a hybrid cloud infrastructure with some workloads still on premises, even as it speeds into the generative AI era, says Nagrath, who became senior vice president of product development in 2020.

“When you’re introducing many new applications, the ease of getting them up and running and lowered costs [on the cloud] is tremendously beneficial,” he says.  

And Nagrath would know, as ADP, with its 10,000 IT pros globally, 70% of whom are developers, started building cloud-native code from the start of its cloud journey. That strategy fully enabled the company’s applications to exploit all the follow-on services developed by Amazon and other cloud providers, and developers in ADP’s innovation labs continue to experiment with cloud-related technologies as they surface today.

While AWS remains a primary partner, Nagrath emphasizes ADP’s use of a range of vendors’ tools to achieve best outcomes. Aside from its use of Azure and Cisco Cloud, for example, ADP has leveraged AWS, GCP, and Snowflake for analytics, as well as myriad AI platforms.

ADP Data Cloud is one of the “richest datasets in the world,” and this enables the company to anonymize, customize, and monetize its data stockpile in many new ways for its client base, Nagrath claims.

ADP combines various datasets and analytics technologies and builds algorithms and machine learning models to develop custom solutions for its clients, such as determining salary ranges for nurses in a specific state that a healthcare client may be evaluating for relocation. Its massive dataset also makes it an industry-leading benchmark for compensation analysis, as another example.

ADP’s journey to AI

As head of product development, Nagrath currently leads ADP’s Workforce Now SaaS, while Global CIO Max Li manages the IT organization.

Together, Nagrath and Li have reached feature-complete status on the digital portion of ADP’s transformation, with AI now the next generation of ADP’s evolution, Nagrath says, noting that the plethora of generative AI tools available in the marketplace has opened a gold rush of innovation.

“There’s a brand-new wave of digital transformation that’s coming,” he says. “Our approach is multifaceted, and we have many initiatives under way.”

ADP’s innovation lab has already developed many machine learning models and predictive analytics that exploit the company’s data cloud. Adding generative AI to the mix will expand those capabilities and provide automatic guidance to employees and managers.

As with the cloud, ADP has partnered closely with AWS, for its Bedrock generative AI platform and its $4 billion partnership with AI startup Anthropic. Still, ADP’s long-term experimentation with AI also includes use of Microsoft’s OpenAI Service and Databricks’ AI platforms, Nagrath says.

The first major product launch of that effort, ADP Assist, was launched at the company’s annual conference late last month.  

ADP Assist combines decision-making with data-driven insights and “streamlines day-to-day tasks for practitioners, managers, employees, and executives with simple, human-like conversations,” according to the company.

The generative AI offering, which is available to a limited number of clients, validates payroll data and identifies missing tax information. Its Smart Analytics gives clients extensive global compensation and turnover data as well as HCM insights. It also features capabilities aimed at commonplace work issues, reducing irritation of employees and handling low-value tasks for HR professionals.

For example, Assist provides customers with call summarizations and applies generative AI to automatically provide effective solutions to problems. “A client may call and tell us that they just hired a person living in Texas and [ADP Assist ] advises the client that they have to get registered to do business in Texas if they’re going to hire there,” Nagrath says. “It’ll tell them they will end up paying penalties later because they are not registered and authorized to do business in that state yet.”

The development chief anticipates far more powerful solutions but faces key challenges as ADP’s ambitions grow. Although ADP has a good number of data scientists, it is challenging to find enough for the company’s goals.

Another challenge is increased competition by the arrival of AI — a powerful tool for novice competitors.  “One major challenge has been the speed and pace of new offerings and new innovation,” Nagrath says of the growing AI arms race. “It allows others to move very quickly as well. We are not the only ones who can move quickly now.”

And, despite how evolved ADP’s strategy has become, Nagrath acknowledges that it is a challenge to plan next steps because of all the unknowns that generative AI poses.

“There are a lot of tools and offerings out there, but it is still so nascent. It is hard to have the ROI and know the efficacy of these things,” Nagrath says. “We are so early in the game and doing a lot of experimentation. We are still forming [a plan] on how we’re going to do it.”

3Sixty’s Tiliakos sees generative AI propelling ADP’s ambitions to the benefit of its customers.

“We’re only in the beginning of the ability to apply generative AI to the ADP Data Cloud to connect the dots,” he said, adding that the implications of generative AI insights to 401Ks and performance management will upscale overall business performance. “This will gradually augment all of our roles and help us to be more effective.”