The challenge

Boasting a number of parks and open spaces, the London Borough of Redbridge is recognized as one of the greenest regions of the Greater London area. In addition, over the past several years, the borough has developed a reputation for innovation and improving the lives of its citizens.

"We are focused on driving change in our community and getting things done for our residents," explains Emeran Saigol, Head of Change for the London Borough of Redbridge. "So we were looking at small scale problems that we could address that would have a direct impact on our customers, and that's when we looked at our contact center."

Every year in the weeks following the Christmas holiday, the center would experience a spike in call volumes related to garbage collection. Given the marked increase in waste and recycling during this period, the borough was forced to increase the frequency of its collection efforts, which often changed the pickup dates for residential neighbourhoods.

"They'd be calling to ask, `When are my bins going to be picked up?'" adds Saigol. "We wanted to take the heat out of the contact center, but the Christmas bin collection schedule was due to start in a 17-day window. We needed something in a hurry, as any solution would have no value going live five days into the Christmas Holiday period."

The solution

That's rubbis.

Moving quickly, the borough developed an AI-based agent able to answer the public's queries regarding waste collection.

"We called it `Binbot' because `Christmas Bin Collection Date Bot' is a mouthful," adds Chris Patterson, Managing Director at EscalateAI. "We had to act aggressively to have it done by Christmas since it's not very useful having a Christmas Binbot in January. From our first conversation to go live was only a two-week period."

Built using the IBM watsonx.orchestrate solution, Binbot allows citizens within the Redbridge region to ask questions regarding collection schedule. And thanks to the Tone Analyzer service in watsonx, the agent can monitor the conversion and escalate it to a call center representative depending on the tone of the discussion as well as the confidence rating of its answers.

"Generic chatbots are good with straightforward questions," explains Patterson. "But as soon as there is any emotion or difficulty — if a user is getting upset or has a really complex question — then a human agent needs to become involved."

The Binbot quickly proved to be a success, leading the borough to examine other areas where it might use watsonx technology to help protect the beauty of Redbridge. "Politically, clean streets are one of our administration's big agenda items," clarifies Saigol. "And the movement of waste collection dates generates a huge amount of contact for the council. But by ensuring residents are able to find out what their collection dates are, we can ensure that we have a clean borough and happy residents."

"After all," adds Patterson, "who wants uncollected rubbish outside their front door?"

Expanding on the Binbot solution, a new visual analysis component was added to the tool, taking advantage of visual recognition service in watsonx offering. The team trained the solution with over 3,000 images of rubbish, enabling it to identify commonly dumped items. Now residents can use their cell phone to take a picture of public waste, such as a rogue mattress, loose bin bags or an abandoned vehicle, and notify the borough by uploading the image through the agent.

The results

"What's really important for us is the value and output," clarifies Saigol. "For a local government organization to commit to new types of technology in times when budgets are extremely tight, we need to show results. Fortunately, this solution did that."

He continues, explaining: "The contact center experienced a 25 percent reduction in calls compared to the Christmas period of the previous year. And on one day, we had about 1,600 conversations through the virtual agent. That level of engagement helps us to be more efficient and can potentially save us money in the long run."

The public also appears to be pleased with their interactions with Binbot. "We built a feedback system into the agent," adds Patterson. "And I would say about 95 percent of the feedback was excellent. They were saying, `Fantastic work. Great agent!'"

Beyond the Christmas season, the rubbish visualization component helps keep streets clean and empowers the borough to be more responsive to the needs of the public.

"Before, we had reports that came in secondhand, maybe five or six days after the rubbish was dumped," adds Saigol "But now it's much easier for people to tell us what is going on out there. We can let our enforcement teams be intelligent in how they go about planning their day-to-day operations. With watsonx, we can be far more proactive and stay ahead of the data trends and as such demand."

About London Borough of Redbridge

With its administrative functions hosted in Redbridge Town Hall in Ilford, the London Borough of Redbridge oversees a host of public services including adult and children social care, regeneration, parking, waste collection and park management. At present, the borough encompasses a population of over 305,000 residents.

Customer stories on this site illustrate what's achievable with IBM-powered Agentic AI. Quoted language and metrics come from public sources or are shared with consent.

← Back to all case studies