Agility’s Humanoid Deployment Process

Our methodical approach to successful customer adoption

Article Sidebar

Introduction

After several years of deploying humanoids in live customer operations, we have a well defined picture of what real integration entails – and a customer program tailored to ensure success. It’s only possible to build a program like this, that’s designed to integrate a new piece of AI powered industrial automation into a brownfield site, when you have actual experience on the facility floor. 

First engagement 

Our first conversations with prospective customers elucidate what sort of workflow they’re looking to automate, and whether a humanoid is even the right answer to that problem. Is the work repetitive? Does it require moving objects between “islands of automation”? Does it fit the classic three D’s of robotics – jobs that are dull, dirty, or dangerous?  

Sometimes, the workflows that customers describe don’t fit a solid use case for a general purpose automation tool like a humanoid. An AMR or a fixed arm may be more appropriate. In which case, we point these prospects to other solutions better suited to their needs.

If the answers do line up during our initial screening, we progress to a bit of table setting on what it actually entails to bring a humanoid into an existing ecosystem. Agility has formalized this process into a tested Customer Acceleration Program (CAP) to identify the use cases that would most benefit their business goals.

CAP process

Here are the three steps we use to deploy humanoids:

1. Solution setup

During our proof-of-technology (POT) phase, we first work with you to define the “win” for your organization – exploring the fine details of potential applications and mapping out exactly where Digit fits in your production ecosystem. Then, we validate a first use case in a controlled environment at our own facility. 

The POT phase looks different for every customer as we take into account what your specific infrastructure looks like and the nuances of your layout. Are there buttons that our humanoid, Digit, would need to push? What obstacles will it have to avoid? What are the weights and form factors of the goods that Digit will be manipulating? 

We recreate these conditions in simulation and physically onsite at Agility to fine-tune the workflow execution. Then, we start generating data from the physical workflow. Beyond what we are able to observe in simulations alone, this real world embodied data helps us refine Digit’s ability to execute in the most efficient way. 

2. On-site validation

Next is our proof-of-concept (POC) phase. After we’ve scoped the workflow at our own site, it’s time to bring Digit to your facility and stress-test the application in your actual environment. We cordon off a workflow site to separate the pilot phase from your official manufacturing line or live floor, and begin to integrate Digit into your ecosystem. 

To get more granular – Agility deployment engineers uncrate the Digits needed for your workflow, connect them to your WiFi network (with the requisite cyber and data security measures accounted for), connect the humanoids to our fleet management software, Arc, build a map of the area that the Digits will be working in, specify where items need to be moved or manipulated, assign the robots their workflow, and then get them to work. 

But here, of course, is where the details of real deployments are essential to acknowledge. When you first integrate any piece of automation equipment, it does not work as seamlessly as you want it to. Adjustments will inevitably be needed that account for the quirks of your facility. Which is precisely why Agility deployment engineers are on site to ensure that these adjustments are made from the start. 

But our attention to deployment realities goes even further than that. 

Considering change management for our customers’ employees helps implementation go smoothly. This means first helping customer management form messaging about the new automation hardware that’s walking around on site. It means training an on-site customer team on how the humanoid and fleet management software operate. And it means providing customers with a framework to address any concerns that arise about our human form factor automation solution. It’s here that we encourage customer management to stress Agility's dedication to safety. Workcells containing Digit have passed OSHA-recognized Nationally Recognized Test Lab (NRTL) field inspections - which helps to ease any reticence around working alongside something as novel as a humanoid. 

3. Operational impact

Lastly, we officially move Digit onto your production line, measuring uptime, throughput, reliability, and operational impact to ensure the solution is ready to scale. By the end of the CAP pilot process, you will have 90 days of operational data quantifying the robot’s return on investment. Experience that will help you decide how to scale Digit across additional workflows, shifts, and sites. Customers are then able to move to an ongoing Robots-as-a-Service (RaaS) model – a CapEx-light adoption method. 

But again – details of real deployments are important to note. As with any industrial-grade automation tool, our humanoids will eventually need repairs, software updates, and ongoing monitoring of the cloud platform that reports on statistics and fleet health. This is where our service and support infrastructure is essential. Agility has built a vertically integrated deployment ecosystem: hardware, software, and the infrastructure needed to ensure that it’s successfully integrated into customer facilities with wildly different use cases and needs. Without this in place, you are left with impressive videos of humanoids doing backflips, but no mature assessment of what it takes to deploy in the field. 

As a final note on deployment safety – for now, our 4th-generation Digit stays in a sectioned-off workcell during the CAP program and during full RaaS deployments. However, we are on schedule to deploy our first cooperatively safe humanoid in 2027, capable of working alongside customer employees without safety barriers, unlocking a new massive potential to scale.  In tandem with this new release, we are diligently helping develop ISO safety standards for humanoids and other dynamically stable industrial mobile robots that will cover this unprecedented technology, ensuring that our customers have a clear safety and compliance framework that lets them trust our solutions.

How your facility can prepare

Not every facility or operation we talk to is prepared to deploy a general purpose humanoid into their worksites today. But if you’re considering what it might take, here’s a short list of what you can consider to get the conversation started:  

Workflow audit readiness: Have you mapped your dull, dirty, and dangerous workflows and quantified their cost? If not, start there. We can't effectively find your best use case if you can't show us where the pain is.

Infrastructure baseline: Do you have proper WiFi and internet connectivity in the relevant facility areas, floor mapping, and an understanding of existing automation "islands" to connect?

Internal champion identification: Who owns the deployment at your site? Facilities? Operations? IT? The pilot goes smoother when there's a named owner before Agility shows up.

Change management groundwork: It’s never too early to begin the internal conversation about automation before the robot arrives. The companies that struggle most are the ones where workforce communication is reactive.

Conclusion

Humanoids have the potential to change the nature of labor, giving you a tireless partner to achieve your operational goals. But those aspirations must be met with attention to detail, diligent safety frameworks, and experience with the frictions of integration. We pride ourselves on remaining an honest voice in the burgeoning and incredibly exciting humanoid market, and believe that our methodical approach to deployment is getting us one step closer to widespread adoption.