Overcome 6 Blockers to use Mobile Robots at Scale

It can take many years to cross the chasm (Geoffrey Moore) for a new technology. Autonomous robots can technically do a huge number of tasks today, but the rest of the ecosystem has to catch up before they can fully go mainstream.

A few weeks ago, I keynoted the RIA Autonomous Mobile Robot (AMR) Conference, where I walked through a set of customer stories from Freedom Robotics’ largest customer deployments, discussing the gaps and blockers which are holding back the manufacturing and logistics industries from significantly increasing the automation and lowering cost for moving things using robots.

In this talk, which you can watch here, I laid out the 6 key blockers I have seen repeatedly and the steps our team at Freedom has taken to enable these companies to massively increase their efficiency and overcome the current chasm of robot deployments today.

RIA presentation about crossing the chasm to full AMR deployments.

Save $1M per robot you buy

Through a set of our customer’s calculations, over the 10 years after a mobile robot is deployed, they save at least $1–1.5M through lowered labor cost and higher efficiency. The numbers add up at scale…

But… there is a huge gap in ability to follow through

Almost every large manufacturing company is assessing and creating a plan to use robots in the next 5 years in larger ways. But, only 1–5% of tasks which robots can do today are automated (Gartner) and just 11% of companies which automate a part of their company with robots succeed at automating a second part (BCG).

Today, robots can easily do 20x more jobs than they are installed at.

The Chasm: 6 Key Blockers for AMRs

• Non-scalable Deployments — Many times, the second and third deployments take as much time/effort/cost or even more than the first one, blocking the ability to increase investments quickly. By designing for incremental scale, companies are able to replicate their first success faster and cheaper.
• Unsharable Tribal Robot Knowledge —Usually, one or two people in a 10,000 person company have the knowledge of how to provision and manage each robotic platform. By investing in team- and company-wide training, the knowledge moves from a few Gurus to being infused into daily work.
• Missing Tools for Humans — Most of the technical tools for robotics were designed for roboticist, so if anything goes wrong, an expert must be called in. If tools are designed with the daily user (not the programmer) in mind, it gives them the ability to act quickly and fix problems themselves, or report the issues with confidence to others.
• Brittle Robot coordination — Most robotic installations were built by a “Systems Integrator” who implemented it then left. If another robot is needed, or if the system has to work with many platforms — ex: UR, OTTO, SeeGrid and multiple PLCs — there is not a standard way to make them all work together. Usually something like Kepware or OPC-UA is used to bridge between them, but it only describes the communication protocol, not the ability to actually coordinate and have robots or other systems work together efficiently.
• Unquantifiable ROI — Many times, once robots are installed, because the actual performance of the robots is not measured, it isn’t clear where the savings are. By creating standard metrics from the beginning which identify the performance, uptime and relative value to human labor, teams can prove their value and identify valid additional investment.
• Lack of Continuous Improvement — Robotics is new. Therefore it never goes perfectly the first time. The challenge is that most industrial installations are done as one-off projects by SI’s. They do their best, but because insights to improve it might not be known for weeks or months after it is installed, the ability to iterate as we might normally do in a standard development process can take quarters or even years per change. By empowering the direct managers of the product with the tools, processes, data and responsibility to continuously improve at a rate measured in weeks and months (while still having valid change control), the actual optimal efficiency of a robotic process can be approached quickly.

I would suggest listening to the talk, where I unwind each of these in more detail. We have see both exciting advancements but also painful examples of very large companies becoming stalled internally by each of these, which has cost them $B’s in efficiency.

Reflections — Cross the Chasm

If you have read Geoffery Moore’s “Crossing the Chasm”, you have heard the argument that all technologies go through an adoption chasm. Autonomous robots are sitting in this chasm today. The tech nology is robust, the use cases are well proven, but it still feels like cell phones in the early 90’s. You knew someone who had one, but most people never used one themselves.

It took 17 years after Cell Phone networks launched successfully for normal people to buy one as a standard technical product.

Robots, and AMRs in specific, need to have subtle improvements to the ecosystem, support, training and other system. Once the world has rolled these out, I am very bullish on even faster adoption.

From working with hundreds of robotics companies, we have seen a set of key triggers for a company successfully adopting robotics at scale. Many of these match the transition from old-school IT to modern day agile cloud technologies.

As robots become more mainstream, the secondary systems which have made Cloud, Mobile and other technologies flourish are coming online. For the companies which embrace and push for a repeated cycle of measurement, continuous improvement and designing systems which are meant to scale, they will do very well in the next decade.

Originally published at http://hansclee.blogspot.com on November 20, 2020.