Tested within the Re_source 3.0 Open Innovation Program, the pilot between Algar and NILG.AI showed how artificial intelligence can cut location-planning time for recycling bins, delivering proposals in hours instead of days.
To meet the ambitious national recycling targets in Portugal set by PERSU2030, Algar, the waste management company serving 16 municipalities in the Algarve, must more than double the number of recycling bins on public roads by 2030. Achieving this means increasing the ratio from one recycling bin per 117 residents in 2023 to one per 46 residents. But until recently, deciding where to place each bin was a slow and complex process, taking up to a full day per location and with a rejection rate of around 30% from municipalities. Within the Re_source 3.0 program by Sociedade Ponto Verde, Algar partnered with the Portuguese strategic AI consultancy NILG.AI to develop an artificial intelligence tool that streamlines site selection for new recycling bins. By integrating rules for placement, such as proximity to homes, access for collection vehicles and absence of obstacles, the tool aims to cut proposal preparation time by 50% while increasing approval rates.
The pilot brought together key actors: Algar provided operational expertise and historical collection data, NILG.AI designed and trained the machine learning model and the Municipality of Lagoa served as the real-world testing ground.
This team developed a model that drew on internal (collection history, current bin locations), open (population and housing data) and external (satellite and Google Street View imagery) data sources in order to predict waste generation based on surrounding points of interest and demographic patterns, ranking potential sites by estimated waste output and service reach.
In the Lagoa pilot, the AI produced a shortlist of fourteen recommended locations in four hours, and the process, which would normally take a technician ten days, was completed in just two. Field verification confirmed a 70% match between AI suggestions and real-world feasibility, with three sites immediately validated for installation, one of which already had a bin not captured in the database, and therefore identified ways to refine the visual model for greater accuracy. The time savings were striking, as the entire proposal for ten bins was finished 83% faster than the manual process and early results also suggest a potential 30% boost in municipal approval rates, thanks to stronger data-backed justifications for each site.
“All of this will contribute to optimizing the selective collection process, not only in terms of accessibility to the population, but also in terms of the amount of waste collected, potentially reducing costs due to a more efficient collection process, and also reducing emissions associated with waste collection. This will bring us closer to a more sustainable future. We all have to do our part.“ – Miguel Nunes, Technical Support & I&D Head Manager at Algar
Looking ahead, the AI model can be adapted beyond siting new recycling bins. Future uses include optimizing collection routes based on predicted fill rates, relocating underperforming bins and even determining locations for biowaste containers. The flexibility of the approach makes it suitable for other municipalities in Portugal or anywhere in the world, where accessibility and efficiency are critical to boosting recycling performance.
“Our role in this pilot was to ensure that Artificial Intelligence was not just deployed as a technical tool, but embedded strategically to deliver measurable value. By applying our AI adoption methodology, we supported Algar in validating how advanced data sources and algorithms can accelerate planning in waste management. The results show what happens when AI is approached strategically: faster execution, more sustainable cities, and scalable processes that can extend beyond waste management. It also sets a precedent for how municipalities can adopt AI responsibly to improve decision-making and sustainability outcomes.” – Pedro Serrano, Data Scientist at NILG.AI
By combining data science with operational know-how, this pilot shows how technology can accelerate progress toward circular economy goals. Meanwhile, for the Algarve, it means faster deployment of recycling infrastructure, greater accessibility for residents and a better chance of hitting PERSU2030 targets on time.
Original Article here