Automatic Data Processing (ADP) Is Set to Win in “AI-Enhanced Human Capital Management” (2 of 2) (Tech Strategy – Daily Article)

In Part 1, I went through the basics of ADP’s business model historically. This was for the decades when it was mostly a labor-focused service business – that then grew in geography and services. That article was a tee-up to this article, where I go into tech strategy.

Recall, my point was that labor-based enterprise services are mostly a long play for economies of scale. That’s what ADP did. And their global scale in human resources and payroll services created two important tech factors. They are:

  • Point 3: ADP’s economies of scale in services enables economies of scale and switching costs in software.
  • Point 4: ADP is a highly integrated “necessary cost”. This may be changing.
  • Point 5: ADP has amassed a massive proprietary dataset on workforce behavior. They are positioned to lead in AI-enhanced human capital management. Their +1M company client list gives them a unique data corpus on how companies are hiring and managing people. This is key for ADP moving into AI-enhanced human capital management.

I’ll go in those points in this article.

Let’s start with ADP’s evolution into a “services plus software” business.

Point 3: ADP’s Economies of Scale in Services Enables Economies of Scale and Switching Costs in Software

As you merge labor-based services with software, you get a couple of important benefits.

First, you are able to tier solutions much better with software added.

You can do lots of versioning with services. You can offer 5 different types of haircuts with the same staff. But software is much, much better at this. Because it has the economics of bits and bytes. Not labor paid by the hour. We can do freemium with software in ways you cannot in labor-based services.

ADP does one step further. They combine software and labor services into different solutions and versions. They can pair human expertise with free software. If you look at ADP’s TotalSource solution (their big package), you can see interesting combinations of software plus human services.

Second, software lets you expand into lower cost markets.

As mentioned in Part 1, ADP now has a really long list of solutions. It’s crazy. Their suite of labor-based services have been combined with software – and this has changed their unit economics. They did a big expansion from large clients to medium and now to small clients. Mostly software-based solutions is how they got to SMEs. And how they got to +1M clients.

Recall, how they present their solutions in their 10k.

Third, you can build economies of scale in the fixed operating and capital costs of IT.

ADP spends about 5% of revenue on “systems development”. Plus, it has significant ongoing capex and M&A in IT. All of this fixed cost spending is about leveraging their big scale into software. This has been their primary strategy for forty years. They moved from the slow economies of scale of services to the fast economies of scale of IT spending.

Fourth, enterprise software is great for creating switching costs.

Doing labor-based human resources and payroll services is good. It creates long-term relationships with clients. But it is not that hard for a client to switch to another provider.

However, installing software is much stickier. Especially when it is on the enterprise side. And when it’s the entire suite of services. That can get you really big switching costs.

A client with the full suite of ADP services looks somewhat like an ERP system. Recall, it can take 6-9 months to implement their systems. And it can take years for multinationals.

I really like the switching costs of enterprise software.

Now, I don’t think ADP is really an ERP system. It’s more like Salesforce. It has lots of integrations with the ERP system and other installed software. The more integrations the better. And ADP has tons of integrations. Here is their description.

Point 4: ADP Is a Highly Integrated “Necessary Cost”. This May Be Changing.

I have frequently written about how business customers view most products and services as “necessary” spending. Think desks and chairs in a school. They definitely need them for their business. But they can negotiate hard with suppliers because they are fairly basic supplies for their business. They always ask for another discount. They switch suppliers frequently. They don’t do long-standing purchase orders. They rationally evaluate the features vs. total costs.

Generally, it is hard to sell “necessary” products and services to business customers. Having switching costs can help. For most, the only strategy is to get dominant supplier power. Dominant pharma distributors like Amerisource do well against big retailers. But this is rare.

However, Warren Buffett has long invested in B2B suppliers for products and services. That’s interesting.

If you look at his investments in companies like Wabco, you see that these products are “critical costs”. Think the manufacturer of airframes for Boeing. Or the explosive pellets within airbags sold to Toyota.

These are “critical” products where the business customers are hesitant to switch suppliers.

  • What if the cheaper supplier’s airframe has problems in five years? Will planes show cracks? Could they crash?
  • What if the cheaper airbag pellets don’t work as reliably? Will we have to do a recall?

When products and services are critical for their product (or for their entire business), they have to consider risk. It’s usually not worth it to take chances. Why switch providers if you only save 5%?

My favorite B2B businesses are those that provide “critical” inputs supplies and a small percentage of the overall product cost. Like the airbag pellets. Why risk the entire company to save 10% on a component that is only 1% of the cost of the car?

Finally, there are “strategic costs”. Where the goal is not to save money. The goal is to outspend competitors on a key activity. Such as in marketing. Or in software development. You don’t want to squeeze every penny. You want to outgun your competitors.

That’s how I think about B2B businesses. They are necessary, critical and/or strategic costs for the customers. And critical and strategic are both good.

ADP has historically been in the necessary costs business. Human resources is a necessary activity. So is payroll. Every enterprise needs it. But if you have a problem in HR, it is not going to crash the company. And you are not going to beat your competitor by outspending them on human resources.

So, ADP is a necessary cost company. And it built some negotiating power with scale. And it built in switching costs by lots of integrations with other software.

That’s not awesome. But they played their cards ok.

However, this may be changing. Because AI-Enhanced Human Capital Management may be turning ADP into a strategic cost business.

Point 5: ADP Has Amassed a Massive Proprietary Dataset on Workforce Behavior. They Are Positioned to Lead in AI-Enhanced Human Capital Management.

Apple’s engineers under Steve Jobs were able to create many amazing products. But the same engineers can’t create anything under Tim Cook. That’s strange.

Why do certain companies have such positive cultures while others seem to have endless internal strife?

Getting human beings to work together in large numbers is remarkably difficult. The variation in performance is very wide. Both between companies and within companies over time.

If you think of humans as the key asset in a company, why is this asset so much more unpredictable than the others? Most all factories work well. Everyone knows how to measure them and get performance up. But workforces are wildly different in their behavior and performance. And nobody knows what to measure or how to fix this.

Recall my digital operating basics. Management and leadership (DOB7) and Teams, Staff and Culture (DOB8xxx) are the most unpredictable. They have the fewest clear methods for improvement.

Yet, ADP says it is in the human capital management business (and payroll services). But human capital management is a chaotic and unscientific field.

But what if AI was able to measure and improve workforces? What if it could make recommendations that really worked and improved human capital performance in a data-driven way?

In fact, human capital management seems like just the kind of question that AI is great at.

AI is very good at messy situations. It is good at finding small factors and correlations that humans can’t really see. AI is also really good at situations where there are hundreds of factors happening at one. Keep in mind, AI can manage ecommerce logistics operations with millions of packages when humans can’t.

Could AI dramatically improvement human capital management?

I think the answer is yes.

But this 100% depends on having huge amounts of data about the workforces at companies. And across lots of companies and industries. This is just the sort of proprietary data that ADP has.

As mentioned in Part 1, ADP has long moved on from “best practices” to fairly sophisticated data products and data insights. ADP has +1M clients in 20 industries and 500 geographic areas. According to its 10k, its DataCloud has 43M employee records, 95M resumes, and 9M job postings.

Its business sits at the center of a ton of workforce data.

  • It uses this data to create insights for clients.
  • It uses this for its ADP workforce employment report.
  • It creates sophisticated skills graphs for companies and industries.
  • It has “candidate profile relevancy” tools.
  • It has an “organizing benchmarking dashboard.”

And this is not just headcounts, turnovers, and labor cost metrics. This is mostly unstructured data about workforce performance and behavior.

At a minimum, ADP’s data corpus creates economies of scale in data. It might create network effects. And, most interestingly, it might enable new AI-services that dramatically improve human capital management.

Last Point: AI Services Are Likely the New Killer Business Model

I love platform business models. The only business model I like more is complementary platforms, which is when several platforms help each other. This is the top of my pyramid.

But I’m coming to accept that a new type of business model is probably emerging. And it may be more powerful. And that is AI Services. Some examples:

  • Uber, Didi and Grab are nice marketplaces for services that connect drivers and riders. These business models have not terribly profitable, but they do have strong defensibility. But I’m pretty sure that robotaxis are going to beat them in ridesharing.
  • YouTube is an audience-builder platform that connects content creators and content consumers. It’s really powerful as a business model. But I’m pretty sure AI-generated content as a personalized service is going to hit them. Either as a serious substitute or a complement.
  • Search engines are arguably the most powerful business model ever created. But ChatGPT is now a serious substitute (or complement) to them.

It strikes me that ADP is very well positioned to go into AI Services in human capital management. And they are talking about this.

ADP is talking about AI for data insights. And an advisor.


Those are good first steps. But I think two longer-term services are really compelling.

  1. AI co-pilots can improve the creation, flexibility, and performance of medium and large teams.

AI is good at tracking and assessing skills. And personalities. And all the soft factors in team formation and performance. An AI co-pilot could help managers find and assemble teams rapidly for new projects. This could be within a company. And within an industry. Think about product managers within companies using an AI service for rapidly identifying, assembling, and managing their teams.

This is about AI-enhanced human capital management. Starting with teams.

  1. AI agents mean workforces will be made of humans and non-humans.

Uber is going to move to a workforce that is part human drivers and part AI drivers (i.e., robotaxis). That will be its fleet in the near future. And we are going to see similar versions of this lots of industries. The human workforce is going to be augmented and/or replaced by a non-human workforce. These AI agents could be AI in the cloud. Or robots walking around the warehouse. Or robotaxis cruising city streets.

How do you manage such a workforce? What metrics do you use? What does human resources and human capital mean for this type of workforce?

I think this is the near future. And ADP is very well set up to move into this space aggressively. Which I hope they do.


Ok. That’s it for ADP. That last part was just me speculating. But I think some M&A is definitely on the horizon for ADP.

Cheers, Jeff


Related articles:

From the Concept Library, concepts for this article are:

  • Economies of Scale: Data advantages?
  • Human Resources / Human Capital Management

From the Company Library, companies for this article are:

  • Automatic Data Processing (ADP)

Photo by Hannah Busing on Unsplash


I write, speak and consult about how to win (and not lose) in digital strategy and transformation.

I am the founder of TechMoat Consulting, a boutique consulting firm that helps retailers, brands, and technology companies exploit digital change to grow faster, innovate better and build digital moats. Get in touch here.

My book series Moats and Marathons is one-of-a-kind framework for building and measuring competitive advantages in digital businesses.

This content (articles, podcasts, website info) is not investment, legal or tax advice. The information and opinions from me and any guests may be incorrect. The numbers and information may be wrong. The views expressed may no longer be relevant or accurate. This is not investment advice. Investing is risky. Do your own research.

Leave a Reply