Learn how to build a business case for condition monitoring of your submersed pumps.

Condition-based maintenance: from technology pilot to implementation at scale (Part 1: Plan)

Table of contents

Used at scale, AI-driven systems that continuously monitor machine health can transform industrial maintenance — but like every new technology, there are hurdles to clear before their full value can be harnessed. There are three critical ingredients in any successful digital transition: people, process and technology. In this four-article guide, we help you create an implementation plan for condition-based maintenance at scale that covers all three.

In this first article, we look at one best practice that will serve you well across all three aspects of your full-scale technology rollout: namely, start with the end in mind.

Start with the end in mind

The first step most companies take toward condition-based maintenance (CBM) is to trial one or more condition monitoring systems on a small scale. After the new tech has proven its ability to consistently identify which machines are degrading and which are not, you can be confident enough to roll it out to the whole company. But you shouldn’t wait until then to start planning your rollout. Even before you launch your first CBM pilot, you should give some hard thought to what implementation at scale will require:

  1. End user buy-in. The success of any technology adoption initiative depends on the support and engagement of your colleagues. You’ll need to communicate the benefits of the new approach in terms that speak to each person’s hopes and frustrations, and meaningfully involve them in the implementation process.
  2. Technology that can scale with you. There are many suppliers offering AI-based condition monitoring sensors and analysis, and it can be challenging to select the right one for the long term. Look beyond the technology itself to the company behind it, and choose one with a track record of successful implementations and a reputation for sticking around to help.
  3. Integration with existing systems. Industrial companies have complex and interconnected systems, and integrating AI-based asset monitoring technology with these systems can be a challenge. You’ll need to ensure that someone has the technical expertise to integrate the new technology seamlessly.
  4. Change management. The adoption of AI-driven asset health monitoring technology requires changes to existing maintenance processes and workflows, which can be challenging to manage. You’ll need to have a change management plan in place that takes into account the needs and concerns of all stakeholders.
  5. Data management. Condition monitoring systems generate a vast amount of data, and your company will own it, even if the vendor collects, manages and analyzes it for you. You’ll need to determine whether to keep a copy yourself for the long term, and if so, how and where you’ll store it.

“One of our customers started thinking about the design of the condition monitoring notifications from the start. They have hundreds of sites and each team has localized workflows for getting data in and out of the company’s central system.”

“The corporate innovation team at headquarters knew that the new condition monitoring technology was going to have to integrate into that existing central system, not enforce the use of its own standalone dashboard. They took that requirement as a filter when considering which technology vendor to choose for the pilot.”


This may sound daunting. The project is going to be more complex and require more thinking up front than you’d probably like. But trust us: the effort to set your company up for successful change is going to pay for itself multiple times over when you take the next steps after the pilot.

Don’t get stuck in pilot purgatory

An initial trial of the technology is a great place to start gathering information for your larger rollout. But this pilot project will in many ways be a non-representative effort: the people involved will be enthusiastic early adopters, the processes affected will be few and local in scope, and the technology used will address a small, carefully chosen subset of equipment where you are confident you’ll see results. The leap from there to company-wide adoption is notoriously huge — so huge it has a name: pilot purgatory.

To avoid getting stuck in pilot purgatory, make sure you plan up front how you’ll get buy-in from widely differing stakeholders, how you’ll integrate the new condition monitoring technology into the company’s diversity of local systems and ways of working, and how you’ll address management concerns about costs, risks and ROI.

Setting up a proper implementation plan from the start will not only help you manage some of the risks the company’s decision-makers might see in committing to a larger-scale implementation; it will let you hit the ground running once approval comes in. So let’s get started!

Start defining your implementation plan

Your plan doesn’t have to be perfect from the start — and in fact, it will never be perfect. Ambitious projects always evolve. Your implementation plan will be dynamic, building in flexibility so that your rollout is resilient to the inevitable surprises down the road. The important thing is to create a strong initial framework at the start. We’ll do just that in the steps below. In later articles we’ll add in specific steps for people, process, and technology.

Round 1: create a strong core

  • Gather a small but diverse team of 5–10 colleagues who are both invested in seeing this project happen and able to provide differing perspectives from across the company. This is your core team.
  • Together, summarize the purpose, scope, and time frame for the project, from the planning stage through full rollout.
  • Starting with any existing information from the project’s strategic plan, brainstorm the concrete outcomes you expect once the new approach has been implemented at scale: X percent reduction in cost, Y fewer hours of downtime, and so on. If the expected outcomes differ per team or site, define them on the per-team or per-site level. Be sure to specify the expected time frame for each outcome, too: will you see this result after a year? Two years? A decade?
  • Don’t limit yourself to strategic outcomes — it’s important at this stage to map out every way that you expect everyday business-as-usual to change for each team. You’ll use all these outcomes in the next set of steps, to make sure everyone’s motivated to see your CBM rollout succeed.
  • With all these points in mind, zoom in on the plan for your initial pilot. If you expect benefit A on a three-year scale, how will that translate to your definitions of success in the intial one-year pilot?

Download the full 18-point checklist

How to write a strong, resilient implementation plan for condition-based maintenance at scale

30.000+ professionals at industrial companies get our insights and best practices delivered monthly to quarterly.

Please select listing to show.