Posts
Tech & Innovation Blog
Putting Technology—and Technologists—First: Digital Transformation at ADP
What We Do, Why ADP, Future of Work
We strive for ADP’s products and services to be universally recognized, easy to use, and accessible anywhere.
Putting Technology—and Technologists—First: Digital Transformation at ADP
Urvashi Tyagi, Chief Technology Officer
ADP has always been more than a payroll company: In addition to payroll software, we also provide thousands of clients worldwide with tools that help them manage HR, benefits, time, recruiting—everything “from hire to retire,” as we say. But while ADP always leveraged technology to do our business, we have historically been a services company. Over the last several years, we’ve genuinely transformed into a technology-first, a products-first organization focused on excellent service.
For ADP, digital transformation is about serving both our clients and our internal associates. To continuously drive value for our clients, we develop the best possible tools. And to create the best possible tools, we need to provide a superior experience for our developers.
Of course, digital transformation is not a new idea at ADP. Modernizing tools and products is a business necessity for any company that wants to stay competitive. But digital transformations often happen piecemeal and in silos—and they often fail to meet their expected results. That’s why ADP embraces a holistic, enterprise-wide approach that relies on collaboration across business units and focuses on cloud technology.
Unifying Cloud Strategy
Cloud technology is vital to our future. It allows us to offer higher resilience, improved security, stability for our applications, and an improved customer experience on the client side. On the enterprise side, cloud technology allows us to access a global infrastructure, simplify our application architecture, and innovate faster, significantly reducing our time to market. So prioritizing cloud strategy is a given—but how do we determine what that strategy should be?
We strive for ADP’s products and services to be universally recognized, easy to use, and accessible anywhere. We’ve had cloud-native applications before, but previously, our developers had to make tradeoffs. That’s one of the aspects of the cloud—you have a lot of options. But that also means the cloud experience can look very different from one developer to another.
To help give our clients a seamless experience of our many products, we want everything to look and feel familiar. That means the way we modernize our technology stack and re-architect products need to be consistent as well, and that’s where a unified cloud strategy comes into play.
When I came on board in 2019, we had different cloud strategies across business units with individual DevOps teams building bespoke tooling for their developers. We decided we needed more product consistency and closer alignment on our principles of cloud strategy, for example, in terms of multi-cloud vs. hybrid cloud. The way we did that was to put our best engineers in the driver’s seat.
Reducing Silos, Aligning Internally
There’s a quote I love from Dan Lyons, the author of Disrupted. He says, “If you want to be a technology company, put the technologists in charge.” That’s what we did to streamline our cloud-based DevOps processes. Technical leads from various business units came together to create a vision and build strategic and technical alignment so we could begin to consolidate. Instead of having 14 independent CI/CDs, we are on a path to two.
We took a similar approach to establish our DevSecOps tooling. In the non-Agile world, there’s a developer, a security specialist, and a System Reliability Engineer (SRE) managing the operations of running your product. With the Agile model, the developer holds all three roles because the operations engineer and the security specialist’s work is now digitized. Most tech companies use DevOps and DevSecOps as a core strategy because it allows the developer to build, secure, and deploy the code and own it—from the dev box into production. This approach leads to better quality and faster delivery.
When we began driving the change to establish enterprise-wide DevSecOps, the chief architect on my team worked with a lead developer from each of our business units, including the Global Security Office. They met weekly, sometimes daily. Each of those groups had its own DevSecOps processes already, but they came up with a unified approach that made sense. The chief architect presented a proposal to start with, and the conversation continued from there. We ended up with what we call an application security workbench. When a developer checks their code into a branch, this tool automatically runs and lets the developer see any security issues in their code and gives them guidance on how to fix them. Further integration of the tool into the Integrated Development Environment (IDE) allows developers to see security issues as they write code and address them in real-time.
Another way we collaborate internally is code sharing across the enterprise. So if you’re building an enterprise product, anybody within the company can look at your code and make changes. That significantly helps with minimizing silos, because now when a team wants to build a new capability, they make it for the whole enterprise. So if you work on enterprise, essentially, the entire company uses what you’ve created in all our products.
Continuously Driving Value with NextGen Products
When developers have an aligned approach to using NextGen tools and technology, they’re empowered to create NextGen products. To that end, our future products will all be cloud-native and embedded with AI and touchless technology. And we’ve already begun adapting some of our current-gen products to be touchless. For example, early in the pandemic, we worked quickly to make our clock-in systems touchless with facial recognition and voice commands.
We know that our business users spend 20–30% of their time looking for information, so we’re working on ways to optimize data analysis. For example, our ADP DataCloud team is a powerhouse for employee-related data. We’ve released several services focused on autonomous analytics in the last year, converting the tremendous amount of data we have into usable insights about how people work.
Another big focus for us is real-time applications, particularly real-time payroll. The acceleration of the gig economy means an increasing demand for payments that can be made instantly at the end of a shift, so creating a real-time payment ecosystem is critical. We’re also looking at blockchain for ID applications.
Ongoing Change, Ongoing Opportunities
There is never really an end to digital transformation. Once you start the work, you have to keep going. To promote continued collaboration and discourage silos, we encourage daily communication between teams, functions, and business units. We know it takes time to absorb and implement change, so we also share information repeatedly and in multiple ways.
At ADP, we recognize that change, like collaboration, is a team effort. When we present our enterprise-wide proposals, we never attach names because our software strategy, design, and development do not belong to any individual. We build off one another’s ideas, allowing us to grow and innovate as a cohesive company.
We’re on a transformation journey and committed to a holistic strategy, which means we are incrementally modernizing. As we do that, we’ll also deliver new capabilities to our clients as we re-architect existing products and update some of our highest-revenue-generating products for the cloud—as well as create new applications from scratch. The possibilities are endless!
Interested in a career in DevOps or DevSecOps?
Visit tech.adp.com to explore tech careers and to subscribe to our blog.
Tech & Innovation Blog
Innovating Retirement: How ADP Uses Machine Learning to Plan for the Future
How We Work, Culture, Team Collaboration
At ADP, people don’t have to be a leader by title. If there is an idea, and you can think big and innovate, that’s all you need.
Innovating Retirement: How ADP Uses Machine Learning to Plan for the Future
As one of the country’s leading HR technology companies, ADP uses its unmatched data in exciting and new ways. We had an opportunity to catch up with two people critical in recognizing the opportunity to innovate and create a machine learning product for Retirement Services.
Hemlata R., Director of Product Development, oversees the entire product development process. In addition to managing scrum masters, architects, developers, and tech leads, she also heads web development, mobile development, and the machine learning strategy for the entire Retirement Services team.
Sanjay V.R. is the Lead Application Developer and oversees the machine learning practice within Retirement Services.
We asked them how their small team creates cutting-edge technology to build data-driven solutions for their customers, and here’s what they said:
First, let’s hear a bit of what brought you to ADP.
Sanjay V.R.
Sanjay: I started at ADP as an intern while I was attending school in upstate New York. Once I completed my internship, I actually had multiple offers to join other companies. I chose to stay at ADP because getting good opportunities is one of the most challenging obstacles in today’s job market, and at ADP, if you put in the work, getting rewarded is the easiest thing.
Hemlata has been my director for 80% of my career, and I’ve been able to turn to her if I have an idea or if I want to pick up a new role or responsibility. She’s always encouraged me. My senior leaders make sure to recognize me for my hard work. I’ve been promoted three times in my three years at Retirement Services, and that speaks volumes.
Hemlata: I also had several offers when I was looking for a change after my last job. I was attracted to ADP because I’d heard that it was moving toward being more of a technology company that valued innovation—and that its leaders prioritized diversity and inclusion. I’ve seen first-hand that you don’t have to have an impressive title to be a leader here. You can be a leader at any level. You can innovate at any level, and ADP supports and invests on that front. I’m so happy and thrilled that everything I had heard about ADP turned out to be more than true.
Speaking of innovation, tell us about the Retirement Services product you built.
Sanjay: People Like You is a new feature based on machine learning algorithms; it helps participants better prepare for their retirement by offering benchmarks on how people similar to them are planning their retirements. For example, we can show you what percentage of your coworkers are contributing to their 401(k)s and how much of their income they’re contributing. Maybe you contribute 5%, and when you see that your peers contribute 8%, you have the confidence to invest more.
In the retirement industry, advisors usually group people by age or salary and then start giving advice. We wanted to answer the question better and offer advice based on what others in similar socioeconomic situations are actually doing.
Hemlata: ADP pays one out of six Americans; the amount of data we possess is unparalleled. When I joined the company, we discovered that many of our clients’ employees do not contribute to 401(k)s. Since we work for Retirement Services, we saw this as a problem. People often look at their peers and follow them, so we asked ourselves how our data could help create a solution.
How did you go about building People Like You?
Hemlata: We tried to combine the mind and the machine by leveraging our experts’ expertise at ADP and machine learning.
Sanjay: We have folks at ADP who have over 20 and 30 years of experience in Human Resources and Retirement Services. As much as data is our strength, our people and their expertise are equally valuable. So first, we talked extensively with our internal stakeholders since they already know the ins and outs of the industry intimately. Then we conducted market research to understand people’s motivations and concerns better about retirement investing.
After that, we went back to our data sets—everything we have from our payroll and retirement resources—and we started looking at this socioeconomic information to see any relevance between multiple parameters. For example, does age or compensation influence your retirement decisions? What if you’re married, single, or have kids? Based on our internal and external research, we were able to identify somewhere around 30 factors that make an impact; we then narrowed those factors based on the extent of their influence on an individual’s decision. Once we started analyzing that data and built models to create the personas, we realized that we had something worth integrating with our existing retirement products.
When we began this project, it started on a small scale. It was just one other data scientist and me. The two of us created the machine learning part of it, but as we built specific pieces of code for the APIs, we pulled in engineers as we needed them.
Were there any complications you had to work through?
Hemlata R.
Hemlata: The tricky part for me was to make sure that we were compliant with all the security olicies. People trust ADP. It’s our brand. That’s why they come to us for payroll, compliance, workforce management, legal, and security solutions. ADP knows what to do and takes excellent care of its customers, and we take this to heart and always obtain the consent of our clients and employees before we include their data. We’re extremely careful to keep all the data anonymous and not look into any specific client or individual employee data.
Sanjay: Yes, ADP is very sensitive toward privacy laws, so we were very specific about reading only as much data as people were comfortable with. One of the biggest advantages we had was that we partnered with ADP’s DataCloud team. They acted like a data custodian in the project and were responsible for making the data anonymous. They also made it possible to identify an employee—only with their consent—if I needed to access that data to connect specific pieces of information.
I’m a millennial, and I’m one of those people who always clicks on “Do Not Sell My Info” on websites. So, I’m particular about my data, and I think I always had that in the back of my mind. DataCloud made my job easy in that regard.
How do you think machine learning will affect your future work?
Hemlata: We are looking at leveraging this concept of combining the mind and the machine on other aspects of our business, such as compliance processes. As of now, we have used descriptive and prescriptive analytics. Next, we are planning to use predictive analytics to help our clients predict the upcoming required actions. ADP and our clients can solve any predicted problems upfront. We’re always trying to see how we can take our ideas and solutions to the next level.
Sanjay: This is the beginning of an entirely new way of thinking about improving our clients’ experience. We want to look beyond traditional solutions to ensure our clients and their employees feel empowered by our products. ADP also has a general excitement to identify pain points to be resolved and processes we can enhance using machine learning.
Speaking of your customers, do you see any results from People Like You? Are more people signing up to contribute to their 401(k)s?
Hemlata: The results are way better than what we expected. Employee contributions and new enrollments have definitely increased. We also saw this product gain so much attention internally within ADP that several other teams contacted us to see how they could leverage similar solutions within their departments. It’s been fascinating to see the outcomes and the interest from all the other teams.
Sanjay: It’s funny because a bunch of my peers was like, “Oh, I don’t really need a 401(k). I’m too young for that.” Then, two or three months after we released People Like You, someone remarked during lunch, “Hey, did you know that I just signed up for my 401(k)?” Then others joined in—four people also signed up. It’s just a wonderful experience when you hear people say your solution impacts their lives.
After we launched, Don Weinstein pinged me on Webex Teams and said what a great job I’d done and that he was looking forward to what I’d build next. It was a total fanboy moment for me.
Hemlata: This goes to show you what I was saying earlier. At ADP, people don’t have to be a leader by title. If there is an idea, and you can think big and innovate, that’s all you need. Once you have that, you can take it to any level, and people will be so open to talk to you, encourage you, and help support any of these thoughts. It’s really amazing to see that!
Interested in a tech career at ADP?
Click here to search for your next move and make sure to subscribe to our blog!
ADP Uses Amazon Neptune to Model the Agile Modern Workplace
ADP Uses Amazon Neptune to Model the Agile Modern Workplace
2020
Global human capital management (HCM) solutions provider Automatic Data Processing (ADP) recently took a bold step as part of its strategy to be on the “innovation offensive.” ADP’s digital transformation unit, Lifion, launched ADP’s next gen HCM, a system designed to model a dynamic modern workplace where work gets done in teams rather than traditional hierarchal structures. But the system’s original database solution used multiple self-managed databases—including relational databases like multimaster MySQL clusters—and wasn’t ideal for completing the complex queries that ADP needed to deliver advanced capabilities to its customers. Plus, the effort of managing those databases was a burden on the next gen HCM staff.
Having already used Amazon Web Services (AWS) for self-managed database provisioning, the company looked to AWS for a fully managed database solution, and it found one in Amazon Neptune, a fast, reliable graph database that makes it easy to build and run applications that work with highly connected datasets. This purpose-built, high-performance graph database engine is optimized for storing billions of relationships and querying the graph with milliseconds latency. By using Amazon Neptune, ADP’s next gen HCM cut costs by eliminating database licenses and increased staff productivity and time to market while delivering customers a unique set of customizable HR solutions.
We like app-level encryption in addition to database-level encryption. When we use Amazon Neptune, the data is already encrypted before it gets to the database, and then it’s encrypted again at rest.”
Zaid Masud
Chief Architect, ADP’s next gen HCM
Using a Purpose-Built Graph Database to Model the Modern Workplace
ADP offers HCM software solutions for automating payroll, core human resources (HR), talent management, benefits, and workforce management for companies ranging from small businesses to global corporations. Since its founding in 1949, ADP has stayed on the cutting edge of HCM technologies, which led to the launch of next gen HCM. “Today there are many types of workplace structures, and traditional HCM systems have not embraced that,” says Zaid Masud, chief architect for ADP’s next gen HCM. The Dynamic Teams features of ADP’s next gen HCM aim to help organizations break out of the traditional workplace hierarchy by taking team members out of silos, improving engagement and performance, and creating a culture of connectivity.
Building such a solution required a database that could seamlessly manage an extensive network of complex data points. Yet ADP’s next gen HCM first launched using dozens of various self-managed databases with a microservices-driven architecture running on Amazon Elastic Compute Cloud (Amazon EC2), a service that provides secure, resizable compute capacity in the cloud. The collection of databases included relational databases like multimaster MySQL clusters, distributed key-value stores, and column family stores. “We have more than 200 microservices,” Masud explains, “and each domain has its own data storage needs. As you can imagine, this grew in complexity and became unmanageable very quickly.”
The company knew it needed a managed database solution to reduce staff workload. It also needed to move away from a relational database to a graph database. Querying relational databases is challenging and required the next gen HCM team to denormalize the data, or add redundant data, in order to speed retrieval, which wasn’t efficient. “When you start trying to ask relational databases complex questions like approval flows, it becomes pretty unwieldy,” explains Masud. In contrast, a graph database offered ADP’s next gen HCM the agile storage it needed. “Graph databases naturally represent your structure in the way it’s designed or visualized, which enables us to build much more dynamic queries.”
The next gen HCM team had a number of requirements for a graph database in addition to a fully managed service. It needed a low-code app development framework that would enable it to develop highly customizable HCM applications without writing code. “We needed to manage people, data, benefits, parallel integrations such as time and attendance, vacation balances, and time off,” explains Masud. “So many things must be customized at every level to account for client-specific needs, regional needs, and compliance needs.” ADP’s next gen HCM also wanted a graph database with open-source standards, which would help the team avoid lock-in. The team first heard about Amazon Neptune when it was announced at the 2017 AWS re:Invent conference. “We felt that Amazon Neptune was a slam dunk because our application was already using these open standards,” says Masud.
Enabling Dynamic Queries with Amazon Neptune
Amazon Neptune easily builds queries that efficiently navigate highly connected datasets, enabling the next gen HCM team to build applications that use ADP’s wealth of data to answer complex workplace questions for a variety of use cases. For example, a company that wants to internally fill an open position can use ADP’s next gen HCM to search for existing employees who satisfy the skill sets and requirements for the role. “Writing a query, plugging in criteria, and viewing a list of employees who qualify are things that Amazon Neptune is very well suited for,” says Lucky Jain, engineering manager, next gen HCM. “It can answer these questions and quickly return the data.” Users of ADP’s next gen HCM also can access a range of capabilities such as data reporting based on specific use cases and criteria-based authorization. This enables customers to build their own teams, or groups of employees, and then create authorization rules for that group, limiting or granting them permission to view information like salaries, personally identifiable information, and business plans. “It has given us broader use cases in our current features for customers that we never thought we could have accomplished in our early products,” says Jain. Key to this customization are next gen HCM’s low-code development platform and Amazon Neptune’s flexible query capabilities. “One of the really powerful things about our low-code app development platform is that it enables you to build no-code graph traversal queries,” adds Masud. “That’s what we use Amazon Neptune for.”
The migration to Amazon Neptune decreased next gen HCM’s total cost of ownership, eliminating the need to have skilled people operating database clusters 24/7. Now staff can focus on cloud infrastructure and site reliability engineering operations, enabling ADP’s next gen HCM to further grow its platform without adding additional staff. “We were self-managing everything from operating system–level things like patching, backups, and point-in-time restores to security vulnerabilities,” says Masud. “Spending less time on those things significantly improves our time to market.” ADP also avoids paying for database licensing and reduces spending on Amazon EC2. Amazon Neptune provides high availability using a minimum of two nodes compared to three with next gen HCM’s former solution. “We expect an increase in reliability and availability with Amazon Neptune, which means that we’re running less of an exposure risk.”
Amazon Neptune has multiple levels of security, including encryption at rest, which is important in securing next gen HCM’s sensitive data, such as personally identifiable information. On an encrypted Amazon Neptune instance, data in the underlying storage is encrypted, as are the automated backups, snapshots, and replicas in the same cluster. “We like app-level encryption in addition to database-level encryption,” says Masud. “With Amazon Neptune, sensitive data is already encrypted before it gets to the database, and then it’s encrypted again at rest.” Amazon Neptune also satisfies end users’ requirements for compliance with Service Organization Control and General Data Protection Regulation. “We’re very comfortable telling our customers that we are on AWS,” says Masud. “They can even register and download AWS Service Organization Control compliance reports on their own.”
Optimizing Workplace Management
ADP’s next gen HCM is exploring multitenancy using Amazon Neptune to better represent the structure of its customer data: currently, each customer has its own isolated graph. Masud says AWS has been very responsive to ADP’s needs: “AWS has a way of getting things out to market quickly and then refining that and iterating over that. We were able to give some direct service feedback to the Amazon Neptune team.” The company is also investigating serverless frameworks on AWS.
With the fully managed Amazon Neptune, ADP eliminated database licensing, reduced Amazon EC2 costs, and enabled its team to focus on core business operations rather than database maintenance. But most importantly, it was able to use the purpose-built graph database to power complex queries and deliver to its customers advanced HR applications that it wouldn’t have been able to otherwise.
About ADP
Founded in 1949, ADP designs cutting-edge products, premium services, and exceptional experiences informed by data for HR, talent, time management, benefits, and payroll that enable people to reach their full potential.
Benefits of AWS
- Reduced costs
- Boosted time to market
- Improved staff productivity
- Eliminated database licensing fees
- Secured sensitive data
- Provided Service Organization Control and General Data Protection Regulation compliance
Tech & Innovation Blog
Team effort: Building a better customer experience through cross-functional collaboration
How We Work, Voice of Our People, Team Collaboration
At ADP, every milestone is achieved–and celebrated—together. The work by Sachin Ghag and his team to improve the year-end testing experience for 401(k) administrators is no exception. Along with the Architecture Group, the “Agile Archers” and the “Avenging Explorers” worked across time zones and collaborated with other ADP teams to bring a brand new user experience to life in four short months. Hear from Sachin about how his team got it done.
By Sachin Ghag, Senior Manager, Global Product Development and Technology, Retirement Services
At ADP, every milestone is achieved–and celebrated—together. The work by Sachin Ghag and his team to improve the year-end testing experience for 401(k) administrators is no exception. Along with the Architecture Group, the “Agile Archers” and the “Avenging Explorers” worked across time zones and collaborated with other ADP teams to bring a brand new user experience to life in four short months. Below, hear from Sachin about how his team got it done.
****
At the start of every new calendar year, retirement plan administrators at millions of companies across the U.S. add the same pesky item to their to-do list: year-end testing. To ensure their businesses are compliant with federal law, they must confirm that their 401(k) plans are within a half-dozen or so Department of Labor standards, which cover everything from which employees qualify for a plan to how much they contribute. If administrators discover any issues, they must resolve them relatively quickly.
For years, this process was largely manual. Test results came in a single PDF, and companies that needed to take corrective actions had no online option to track whether those issues had been marked as resolved. Instead, administrators had to call ADP for a status update. But we saw an opportunity to save our customers time and better use our resources. It was clear that we could offer a better experience.
We’re always looking for ways to optimize our processes, so a self-service dashboard for year-end testing had long been on our list of projects to tackle. And the timing was perfect: ADP had just started accelerating digital transformations across the company. Our plan was ambitious: We wanted to give clients not just a real-time view of their status, but a central hub for every resource they’d need to resolve any issues. By September 2019, our team in Global Product and Technology (GPT) was ready to dive in. But we knew that the GPT team would need help from our colleagues along the way.
To kick things off, we held a discovery session with the dashboard’s product owner and the UX team, who had already created a mock-up of the end-to-end user experience. Once we made sure we fully understood what we needed to build, we broke the desired product down into features, created user stories for each one, and developed a timeline based on three-week sprints. We wanted to leave plenty of time to test every scenario before the January 10th launch, so we set a target date of December 12.
The first step, we knew, would also be the hardest: Before we could build the APIs and UI that would make our self-service dreams a reality, we needed to move data from a highly complex mainframe system—which most of the Retirement Services GPT team had never worked with before—into SQL. So once our chief architect had offered some invaluable initial feedback, including new processes for transferring the mainframe VSAM data into SQL, we turned to ADP’s subject-matter experts: the Mainframe team. Together, we decided they would extract the millions of records we needed into a text file, updated daily, which we would then import into the SQL site. And of course, updates had to go both ways; we also needed to figure out how to send changes back to the mainframe—re-running tests as soon as possible after users completed corrective actions to ensure the two sources were synced.
Throughout the development phase, collaboration was key. A challenge with the time difference, when our India Global Product and Technology team was half a day ahead of our Mainframe colleagues in the U.S. Being flexible and thoughtful, we managed to meet jointly for an hour or two nearly every day, and were even able to turn the time difference into an advantage. Because our U.S. teammates worked while we slept, they would often have suggestions and solutions ready for us by the time we started the next day.
Once the initial development work was done, yet another phase of collaboration began. We asked our colleagues in Service Operations, who work directly with clients, to help us test the dashboard. Sure enough, their real-life experience helped them find issues we hadn’t—especially around ADP’s 401k Sponsor site, which is used by plan administrators. If a user clicked on certain links within their ADP Task Tracker, for example, we wanted to send them directly to the new self-service dashboard—but many of those links still needed updating. The Service Ops team recorded each issue they found in a spreadsheet, and we fixed them, one by one.
In the end, thanks to hard work from our team and our colleagues across UX, Mainframe, Service Ops, and beyond, what started as an ambitious plan turned into a success story for our teams and our clients. In early January, we launched smoothly, on time, and with a warm welcome from tens of thousands of happy clients—whose reviews ranged from “I love how easy this was to navigate” to “You made my freaking day!” ADP’s leadership team also recognized our work with an award of appreciation.
The new year-end testing dashboard.
Since that first release in January, we’ve already built out some additional features—and we have plans to add more for 2021, including web identification of data integrity issues, which will allow our clients to visualize and modify data within their web session. But even when we aren’t actively working on the dashboard, the experience of building it continues to benefit the GPT team every day. We’ve been able to use the technical knowledge we gained to improve our work on several other projects, both within and outside of compliance. And most importantly, we’ve built relationships with other ADP teams that will help us better serve our clients for years to come.
Given all the things 2020 has given us, our clients will have a smooth year-end. A nice gift after everything that has happened in the world.
Want to Meet the Team?
Tech & Innovation Blog
Time clocks in the age of COVID-19: How ADP developed a (mostly) touchless solution in just a few weeks
Pandemic, Innovation, Voice of Our People
In early March 2020, as the COVID-19 outbreak started to expand throughout the U.S. and Canada, the team working on ADP’s new Time Kiosk system started getting the same question from many of our clients: “Is there a way to use this without touching it?”
By Jonathon Gumbiner, Senior Product Manager
In early March 2020, as the COVID-19 outbreak started to expand throughout the U.S. and Canada, the team working on ADP’s new Time Kiosk system started getting the same question from many of our clients: “Is there a way to use this without touching it?” At the time, we were several months into a pilot program for the tablet-based timecard management app with more than 1,000 clients.” But most of them hadn’t adopted the app’s facial recognition feature, instead opting to tap in their badge number. And even for those who did use facial recognition, Time Kiosk still required each worker to tap the screen—a suddenly dangerous proposition during a global pandemic.
For companies that had essential workers on-site, we suggested an immediate but imperfect solution: low-cost touch pencils each employee could use to navigate the app. But we knew we’d ultimately need an integrated, turnkey option—and we’d need it as soon as possible.
After a quick brainstorm, we narrowed in on the fix that seemed most promising: First, we’d reconfigure the app to perform facial recognition by default, whenever someone was in front of the camera. Then, we’d use the tablet’s built-in virtual assistant, which powered features like Siri and Google Assistant, to respond to voice commands within Time Kiosk. If we were successful, employees would be able to start a workday, take lunch and other breaks, and clock out, all without touching the screen.
Within a couple of days, our developers were able to build a rough proof-of-concept. It was clunky and far from intuitive—to clock in, for example, you had to say “tap clock in” instead of simply “clock in.” But it was enough to help our senior leaders understand our vision for a more-refined solution—one that would meet the high standards we’d set for the original Time Kiosk experience. We got their buy-in and started to build.
Voice recognition was the first challenge. For one thing, as anyone who’s used a virtual assistant knows all too well, there are phrases it just won’t recognize. Also, in order to release the touchless features as part of Time Kiosk’s formal launch in both the U.S. and Canada, which was just a few weeks away, we needed to develop voice recognition for not only English but Spanish and French, as well—languages no one on the team speaks. Thankfully, as a global company, our partners from other ADP teams came to our rescue, helping us quickly create a repository of words to which the tablets would reliably respond.
Of course, we couldn’t make every action completely touchless. Switching between an employer’s custom job or department codes, for example, would require an employee to scroll through options that voice recognition likely wouldn’t cover. But what we could do was keep people informed. With the help of our UI team, we developed a treatment to add an icon for every touchless function, so employees could see at a glance whether they’d need to touch the screen. If so, they could wash their hands or take other precautions before they acted.
Once we’d finished the first phase, though, we came to a larger challenge: quality assurance. We spent twice as much time testing the new touchless features as we’d spent building them, going through every single action a user could take to make sure we’d identified everything properly. Because voice recognition touched the entire product, we had to review it all—and quickly, requiring a true team effort from QA. What’s more, we happened to be in the middle of transitioning to a new UI, so we needed to test both the current and the incoming interfaces, making the process twice as long.
Yet perhaps the biggest challenge of all was the pandemic’s impact on how it got done. To make sure the new features worked well in all three languages, we needed service reps and tech partners to help us with testing. But most of ADP’s 58,000-person team was working from home. I couldn’t simply walk upstairs, hand someone my tablet loaded with the latest version of our work and ask them to play around with it for a while and bring it back at the end of the day. Instead, we had to find a way to get it on their tablets remotely—no easy task given the tablet’s security restrictions. Luckily, our team was able to build a package and set of instructions that I could share, allowing partners to offer live feedback via an embedded diagnostic tool. They were invaluable in helping us fine-tune, especially our translations.
In the end, thanks to the hard work of everyone on the Time Kiosk team and many of our colleagues, we were able to meet our goal, transforming the app into an intuitive, mostly touchless experience in a few short weeks. Like any quick-turn project, it wasn’t without a few bugs. But the team’s rapid response to client questions and weekly Q&A calls have helped us not only serve their needs and build stronger relationships with them. Time Kiosk has now officially launched, and our sales teams tell us the touchless technology has been a conversation driver with both clients and prospects.
Even after the COVID-19 outbreak has passed, we see great potential in what we’ve learned about voice and facial recognition, whether it’s better accessibility for employees with disabilities or voice biometrics for authenticating service calls. In the meantime, we’re proud to say that when our clients had an urgent need, we were able to quickly deliver a solution that works—and that’s helping keep thousands of their people safe.
Jonathon Gumbiner is a Senior Product Manager at ADP in New Jersey.
Tech & Innovation Blog
How ADP leveraged new technology to help users affected by COVID-19
Pandemic, CARES Act, Helping Clients
When the Coronavirus Aid, Relief, and Economic Security Act, or CARES Act, became law on March 27, 2020, teams across ADP had already been hard at work for weeks preparing for the flood of new policies tied to this legislation. Here’s an example of how Cary Feuer and his team jumped to our clients aid.
By Cary Feuer, Director of Product Management, ADP Small Business Services
When the Coronavirus Aid, Relief, and Economic Security Act, or CARES Act, became law on March 27, 2020, teams across ADP had already been hard at work for weeks preparing for the flood of new policies tied to this legislation. In Retirement Services, we’d started with the simplest—and highest-impact—changes, such as initiating loans and withdrawals for users affected by COVID-19. By mid-March, we had successfully worked through those immediate projects and then turned our attention to a provision we knew would be much trickier: payment suspensions for 401(k) loans.
Since long before the pandemic, the IRS had allowed 401(k) owners to borrow money from their accounts for what it deems “immediate and heavy financial need,” such as a medical expense or a looming foreclosure. Now, under the CARES Act, borrowers affected by COVID-19 could choose to pause payments on those 401(k) loans until 2021. We knew up to 175,000 ADP users might qualify, and their average monthly payment was $800—a significant amount of money for many families. And we also knew that if even 10% of that group decided to suspend their payments and had to call us to do so, it would likely put a significant strain on our team. More importantly, it would be a headache for users during an already-difficult time. We wanted to give them an easy, self-service option, instead of making them wait on hold.
It was clear we needed a technical solution. But speed was critical—and because suspending payments is a multistep process (including self-certification of COVID-19-related hardship)—it wouldn’t be as simple as checking a box. On the backend, we needed to update money-movement databases and multiple payroll products, reamortize the loans, and create an audit trail, all of which we knew we could do relatively quickly. On the frontend, though, we would normally take our time on development and testing, ironing out every wrinkle to ensure the best user experience. A UI build of this scale might take several sprints to ship across mobile, web, and legacy web platforms. In this case, we didn’t have that long.
Instead, we turned to a new piece of third-party technology, which ADP had recently integrated to allow for faster deployment of simple features like pop-up guides and mini-surveys. Designed for product managers and others to use without the help of an engineer, this technology offers templatized, customizable design patterns—and it had already been vetted by ADP’s Technical, Security, and Legal teams. It was our best, and perhaps only, option to get the frontend of payment suspensions up and running on an accelerated timeline. However, because of all the backend changes each payment suspension would trigger, we’d need to learn how to work with the product in an entirely new way, pulling information out of its API and into our own infrastructure.
Our lead developer joined with our lead development team for a quick feasibility study, and within a couple of days they’d determined our plan could work. So, with added help from one of ADP’s resident experts on the 3rd party software, we all got to work building. Our colleagues in Service Ops helped us develop the content, a UX teammate gave the frontend flow their blessing, and in less than two weeks we were almost ready to ship.
But then we ran into a snag. In order for the third-party product to know which users should see a payment suspension option, it needed to refer to a list of qualified users’ anonymized IDs—and with so many people facing financial hardship and taking out new 401(k) loans, that list was changing every day. Because of the time crunch, we’d decided to upload up-to-date CSVs of user IDs to the product each morning by hand. But this seemingly simple fix was a use case that the product—a relatively new technology still in its startup phase—wasn’t built for. Each day’s upload was taking hours to complete.
Rather than delay the release, we decided to ship our new feature and keep handling the CSVs manually. Contemporaneously, we started work on a mini-app that could automatically break up and upload the CSVs. After a few days of testing, we finally had a feature that was not only fully self-serve for our users, but fully automated for us. Thousands of people have now paused their loans without needing to call in, saving them time and potential frustration—and saving ADP the equivalent of adding two full-time employees. Over the course of the program, our uploading solution will save hundreds of additional hours.
Meet Cary’s four-legged office mate
Even better, our team is more familiar with a brand-new technology that we can now leverage in other creative ways. The next time we’re responding to a fast-developing situation, such as a hurricane, we’ll have this 3rd party technology in our toolbox. We’re currently validating it for other use cases, where time to market is less of a concern. With just a few weeks of work, we were able to expand our team’s development toolset, better serve our users when they needed it most, and make an investment in the future of ADP.
This is just one way that our tech teams have added new tools into our tech stack. This feature is now available for all ADP Retirement Services clients that offer CARES Act provisions to their employees.
Cary Feuer is a Director, Product Management for Small Business Services at ADP and is based in New Jersey.
Podcast: Insights into ADP Engineering with CTO Tim Halbur
ADP has been around for more than 70 years, fulfilling payroll and other human resources services. Payroll processing is a complex business, involving the movement of money in accordance with regulatory and legal strictures.
From an engineering point of view, ADP has decades of software behind it, and a bright future of a platform company used by thousands of companies. Balancing the maintenance of old code while charting a course with the new projects is not a simple task.
Tim Halbur is the Chief Architect of ADP, and he joins the show to talk through how engineering works at ADP, and how the organization builds for the future of the company while maintaining the code of the past.
Sponsorship inquiries: sponsor@softwareengineeringdaily.com
Transcript
Transcript provided by We Edit Podcasts. Software Engineering Daily listeners can go to weeditpodcasts.com/sed to get 20% off the first two months of audio editing and transcription services. Thanks to We Edit Podcasts for partnering with SE Daily. Please click here to view this show’s transcript.
