“Can a robot navigate on a construction site and collect information for situational awareness?”
“Does a construction process ontology help combine information from multiple sources to support construction management?”
“Can a digital twin platform serve for visualizing 3D zoning data?”
These are some of the questions our clients have answered through Proofs of Concept (PoCs). I’ve also helped set up PoCs to explore technologies within my own company. But how can we ensure a PoC serves its purpose?
What a PoC is (and Isn’t)
A Proof of Concept is a focused, small-scale, and time-bound experiment that validates whether a technical idea is viable. Its purpose is not to build a working product but to answer a specific question: Can this work?
Instead of launching vague initiatives like “Let’s explore machine learning,” effective PoCs are grounded in concrete goals, such as testing whether Azure Data Services can reduce the time spent producing financial reports.
Unlike prototypes or Minimum Viable Products (MVPs), PoCs do not aim to demonstrate usability or market readiness. Instead, they help organizations decide whether an idea is worth further investment.
When to start a PoC
PoCs typically begin when someone identifies a promising technology or sparks an idea that needs testing. Sometimes, this comes from internal curiosity. In other cases, a customer may initiate the process.
For instance, I once published an article about idea campaigns. A client reached out, interested in launching one within their organization. At the time, I had developed a simple app for collaborative ideation. We launched a PoC using the tool, and over time, that small test evolved into a commercial application used in more than ten countries, including by that same client.
How to set up a PoC
PoCs have a sponsor, typically a decision-maker, and a small, cross-functional team committed to the process. This team may include a project manager, technical experts, analysts, and representatives from relevant stakeholder groups. External consultants or vendors may also participate if their expertise is required.
The planning phase should outline the specific problem to address, the approach to be taken, the expected outcomes, and how success will be measured. Even if the project is low-cost, defining a rough budget helps ensure clarity and accountability.
While PoCs are generally short-term, lasting between four and eight weeks, the timeline can vary depending on complexity.
How to carry out a PoC
It’s vital to ensure that everyone involved has the necessary skills and knowledge to execute the PoC effectively. This might seem obvious, but in practice, it’s often overlooked. I’ve seen tools dismissed simply because users weren’t trained properly.
A PoC doesn’t have to rely on fully integrated systems. For example, if you cannot automate a data connection, manually entering the information can still provide the needed insight. The goal is to test feasibility, not deliver a production-ready solution.
Throughout the PoC, keep track of the progress. Gather both technical and user feedback, and organize midterm check-ins to identify what is working, what is not, and whether changes are needed. Also, be aware of potential risks, such as timeline overruns or unclear roles, that may impact outcomes. Addressing these proactively can help avoid unnecessary delays or failures.
Communication is key
Communication is essential at every stage of a PoC. Before it begins, you’ll need to pitch it to internal decision-makers. This means clearly articulating the business problem, why solving it matters, and what kind of return on investment the PoC could deliver.
During the project, maintaining engagement with stakeholders helps sustain interest and ensures alignment. Once the PoC is concluded, summarize the findings in a concise and accessible manner. Whether in a report, a presentation, or even a short video, the results should be easy for others to understand and act upon.
Some of our clients have used internal blogs and videos to share PoC results and lessons learned. These stories help others see the value of experimentation and build momentum for broader change.
Evaluating and making decisions
After the PoC ends, revisit the original goals and assess whether they were met. Examine how the concept performed relative to expectations and, if relevant, compare it to other possible solutions.
In one case, our client tested a robot on a construction site. While the robot wasn’t autonomous yet, a researcher controlled it remotely and confirmed that it could safely navigate the site, open doors, and collect valuable data. This experience justified further investment into AI-based autonomy, not based on hype, but on objective evidence.
And if the PoC failed, that’s not a failure. It’s helpful to know what doesn’t work. What matters is understanding whether the issue lies in the concept or the execution.
Scaling the results
Even when PoCs are technically successful, scaling them across the organization can be difficult. Often, the new solution doesn’t align well with existing processes, business models, or skill sets. In construction, especially, where collaboration spans many parties, one company’s innovation may fall flat if others aren’t on board.
In some cases, the PoC didn’t deliver clear business value, even though it “worked.” That’s why it’s essential to include business criteria and stakeholder alignment in the early stages.
Still, when PoCs are well-executed and thoughtfully communicated, they often create a lasting impact. They can unlock new ways of working, build confidence in innovation, and mitigate the risks associated with larger-scale adoption. Including external stakeholders in the PoC process can also help ensure broader acceptance and integration later.
A first step
PoCs are a powerful tool for innovation in the AEC sector. They let you explore possibilities, test ideas with minimal risk, and generate the evidence needed to move forward, or not.
With a focused scope, engaged stakeholders, and clear communication, a PoC can become the first step toward meaningful change.
View the original article and our Inspiration here
Leave a Reply