John Sumser: Here’s a view from inside the largest HR tech company

At $14 billion in annual revenue, ADP dwarfs the competition. Its 800,000 customers span the entire range of settings for HR. From the smallest, tiniest niche operation in the industrial center across town to the largest global behemoth, no one understand more clearly the array and complexity of HR execution. The company’s Research Institute produces a monthly jobs analysis that is better than the government’s Current Employment Statistics report.

In September, I spent a day in ADP’s innovation lab in the Chelsea neighborhood of Manhattan, home to large chunks of the exploding digital industry in New York City. The ADP lab is tucked among the other technology giants.
Being the biggest is not always an easy thing. In some quarters, ADP is the company people love to hate. Entire ecosystems have grown up in its wake, and ADP is the target of tons of small competitors who are no more than 10% of its size, mostly much smaller.

ADP serves many niches and tailors its offerings to meet those demands. There isn’t a one-size-fits-all branded solution, as it must serve the needs of all clients, who come in all sizes and many industries across the globe.
There is a deep logic and flow to its product-development process. These days, the process of refreshing and renewing offerings begins in the enterprise toolkit and migrates out to the smaller customers over time.

ADP’s vision and ambition are also big. ADP intends to continue to demonstrate its technical prowess by creating an enterprise toolkit that enables the sorts of client-specific accommodations required to keep 800,000 clients happy. The vision begins with architecture, then imagines ubiquitous AI, delivers business relevant insight to individual managers and even allows clients to create their own tools.

That’s a lot to digest.

The overall software project is a massive trans-national collaboration with design and coding happening in multiple centers around the world. Built on a set of interchangeability standards, the output is a modular design with lots of block-sized components that are melded together in the installation process. The goal is a universal tool that comprehensively meets the needs of individual customers. The distributed development approach reflects the fact that the head of global product and technology, Don Weinstein, is an engineer by training. The outcome objective is audacious.

ADP is taking the view that AI belongs anywhere it can find a home. Their explicit “AI everywhere” strategy takes what other enterprise players are effectively doing and makes it a commitment. “AI everywhere” means that the design and development is looking for places that ADP can leverage its huge data stores and massive client base.
Another principle at the heart of the process is “No Code Development.” ADP wants its clients to be able to easily develop their own applications. The idea is that a customer should be able to point to data sources and lay out logic as a way of developing apps that solve specific problems.

A great example of this is the Chatbot Creator they showed me. With simple logic that you can define as a graphic, it’s possible to create a chatbot that performs the fundamental tasks of repetitive data collection or screening with knockout questions. This will be a much-copied feature as the industry moves towards a purely conversational interface with HR.

See also: ADP rewrites payroll

These bits and pieces are wrapped around ADP’s most interesting application of its data. Their mobile app, the fourth most-downloaded business app in the Apple iPhone ecosystem, can be tailored to deliver relevant business alerts to users. The system allows integration with other key business data to produce alerts and recommendations tied to the key objectives of individual users.

This is where the practical line between AI and big data blurs to the point of meaninglessness. At the core, this tool compares performance against targets then delivers forecasts, alerts, notices and insight. It’s very math-y. In early configurations, the machine-learning components are limited, but that doesn’t interfere with the power of having real-time data immediately available.

ADP is building new AI-based tools that emerge from its core strengths and beliefs. Unsurprisingly, that’s the approach the company is taking across the board. This is practical intelligence that cares more about value delivery than whether or not it is AI.

Leave a Reply