Business talk artifical intelligence : Greyparrot's Gaspard Duthilleul: Why AI in waste management means upskilling, not downsizing
Greyparrot's AI systems analyse waste streams in real-time. How do you train facility operators to interpret and act on this data effectively?
We wanted to make our data accessible and usable with as little training as possible. Apart from data analysts or the like, no one wants to look at spreadsheets or charts. As an operator, you want to use tools that are tailored to your needs. That’s why we spent a lot of time developing intuitive dashboards that make data easy to understand and act on by anyone. Our alert system, for example, uses simple colour codes (Red/Green) that are immediately understandable. Our summary dashboard was designed to replicate the weekly reports that our clients were already generating manually. We also know that our users have different needs. The plant manager, the performance engineer, the maintenance manager and the commercial team all have different perspectives - and the tools they use should reflect that. We train users depending on the way they'll use our data. Our implementation team calibrates the Analyzer system in collaboration with the people who will be using it to ensure they actually understand and adopt the technology. We have also built a knowledge hub featuring videos and explainers, which we are continually expanding with new content to help customers implement the system.
What's the learning curve like for workers transitioning from visual sorting to AI-assisted operations? What support do they need?
With any new technology, it depends on the company’s adoption, readiness, and adaptability. We see some clients jumping on our analytics portals as soon as they’re live and spending hours on them, testing all the possibilities. In other cases - sometimes within the same organisation - we see team members waiting until they are forced to use the system. We’ve learned that the key to adoption is helping people understand how our technology helps them, specifically. How can our technology make their work easier, less hands-on? How does it help them transition to other tasks that they find more rewarding? People need personal incentives, and forcing them to adapt doesn’t yield great results. Support from management, training, and adapting our systems to their needs is critical.
We have six years’ experience implementing AI, giving us deep waste management expertise that shapes how we support implementation. This background means we understand the realities on the ground - the pressure points, the skills required, and the challenges teams face every day. We built this technology to support waste managers in shaping more effective ways of working, it has become a tool for training and upskilling.
Our AI is designed to make facility managers and their teams more effective, not redundant.
How has the partnership with Waste Robotics changed your perspective on the human-robot collaboration requirements in waste facilities?
Waste Robotics are very aligned with our values and our approach to problem-solving. In waste facilities, people bring unique skills - they can interpret insights, manage complexity, and make decisions that drive real impact. Robots, on the other hand, can take on the repetitive or hazardous tasks that put workers at risk. This partnership has helped us better define where people add the most value, and how robots can step in to make their work safer and more rewarding.
Your AI provides detailed waste analytics. How do you help facility managers use this data to upskill their workforce rather than downsize?
Our AI is designed to make facility managers and their teams more effective, not redundant. We use waste analytics to help our customers make informed decisions and to automate what should be automated. Our job is to enhance the workforce’s capabilities, make their work easier, and enable them to achieve more. If you look at some of the roles in an MRF, you might have a maintenance engineer spending a good part of his time trying to identify problems creating machine blockages and downtime, or a performance engineer doing weeks of testing to pinpoint inefficiencies. The real value is in the decisions they’re making based on the information, not in the time spent gathering it. By fast-tracking access to insights, we enable them to focus on their core skills and deliver the most value.
A great example is our work with KSI recycling in the Netherlands. By using our waste analytics, they were able to pinpoint exactly where material loss was happening on their lines. Instead of cutting staff, this insight helped managers train operators to respond faster to changes in the waste stream, reduce downtime, and reallocate people to higher-value tasks like quality control and maintenance. What we’ve seen with KSI is that when workers understand the 'why' behind the numbers, they gain new skills, become more engaged in improving operations, and take pride in driving better results. That’s what upskilling and aligning teams looks like in practice.
What role do experienced sorters play in training and validating your AI systems? How do you leverage their expertise?
We don’t rely heavily on sorters to train and validate our AI. We know our customers are extremely busy, and that our system is mature enough to start providing value without that much input. We have a strong team leading deployment, too - they combine waste sector experience with machine learning expertise. For example, one of our analysts previously led a team of waste samplers at a major facility and now guides the deployment of our solution. Our Head of Data was also an analyst at one of the world’s largest waste management companies. We’ve made it a priority to reduce the amount of work our customers need to do to use our system. With that said, every customer is different. Many of our early clients ran manual checks against our system, which helped build trust in the accuracy of our AI model and improve it. We also listen to feedback: our partners and clients have helped us grow our AI waste recognition library by telling us what materials they need to track. Our technology has evolved according to industry needs.
>>> Using AI to boost Europe's critical raw materials supply
Data’s power is only unlocked when an operator, engineer or salesperson actually uses it. It’s not a switch. Our system is there to eliminate the repetitive, time-consuming work that doesn’t require as much expertise but is still essential - like manual sampling - so that workers have the insight they need to make informed decisions and take meaningful action.
From an operational standpoint, what are the biggest change management challenges you see when facilities adopt AI-driven sorting?
In the past, encouraging reluctant users to adopt our technology was a challenge. That was often down to a lack of trust in AI’s ability to recognise waste accurately. Thankfully, it’s not really a problem we encounter anymore. As the technology has been deployed more widely, there’s a better understanding of its capabilities in the sector at large, and less scepticism about its accuracy. In some cases, slow adoption isn’t down to scepticism - it’s because this data has never been available before. We’re able to gather so much granular insight into waste streams that it can be difficult to know where to begin. That’s why we made it a priority to tailor our dashboards to each individual role and focus on surfacing the insights that matter most.
How do you address concerns from workers who feel their expertise is being replaced by algorithms?
In short, we point out that their expertise is not being replaced. Instead, we’re giving them tools that allow them to get more out of their skills and experience. Data’s power is only unlocked when an operator, engineer or salesperson actually uses it. It’s not a switch. Our system is there to eliminate the repetitive, time-consuming work that doesn’t require as much expertise but is still essential - like manual sampling - so that workers have the insight they need to make informed decisions and take meaningful action.
What organisational culture changes have you observed in facilities that successfully integrate AI systems versus those that struggle?
We see a “no going back” attitude at facilities that adopt our technology. Once teams have access to a continuous stream of data on their material and start making real-time decisions based on actual composition or performance, they can’t imagine their operations without it. It often encourages real curiosity and creativity, too: we regularly have clients brainstorming new ways to visualise the data, or asking for new data points and features. Once they’ve seen AI’s potential firsthand, they start finding new ways to use our system and to operate a more efficient, profitable business.
