Floward uses AI agents to handle seasonal WhatsApp spikes

Floward has deployed Infobip's AgentOS to manage spikes in customer service demand, with the system handling peak WhatsApp volumes during major seasonal surges.

The online flower and gifting company said the setup enabled it to cope with conversation volumes up to 13 times higher on peak days, including Valentine's Day, when it handled more than 54,000 daily conversations. It introduced a group of specialised AI agents to replace rule-based chatbots and route more complex cases to human staff when needed.

The move addresses a persistent operational challenge. As a same-day delivery company serving customers in the UK and the Middle East, Floward faces sharp increases in demand around occasions such as Mother's Day, Valentine's Day and Ramadan, when queries about orders, delivery addresses and changes can rise quickly.

Under the new arrangement, customer conversations are routed to different AI agents depending on the task, including address collection, frequently asked questions and order changes, while live agents remain available for escalation. The service runs through WhatsApp, where recipients can provide address details and ask related questions in the same chat.

Floward said the deployment was designed, tested and launched in less than two months. It reported a 15% reduction in customer service costs and a 12 percentage point increase in customer satisfaction after introducing the AI-based system.

It also said it maintained one-minute response times and achieved 95% service level agreement performance during its busiest periods. AI containment rates also improved from the previous year, indicating that a larger share of customer interactions were resolved without being passed to human agents.

Seasonal pressure

Retailers and delivery businesses have been experimenting with generative AI and automated customer service tools to absorb temporary peaks in demand without increasing staffing at the same rate. The challenge has been particularly acute for businesses tied to seasonal events, where traffic can surge for short periods while customer expectations for speed and order accuracy remain high.

Floward said the project was intended to show that customer service scale no longer had to depend on proportional increases in headcount. It worked with Infobip to redesign customer journeys so more steps in the support process could be handled within a single messaging conversation.

"One of the biggest achievements of this transformation was proving that scaling customer service no longer means scaling headcount at the same rate," said Lujain Mallosh, Customer Care Senior Manager at Floward.

"By combining AI agents, WhatsApp journeys and customer data on a single platform, we significantly improved efficiency and automation while maintaining high service standards during our busiest periods. As a result, we reduced customer service costs by 15%, handled 54,000 conversations on Valentine's Day alone, improved AI containment rates year over year, and consistently achieved one-minute response times with 95% SLA performance. This approach also helped us increase CSAT by 12 percentage points, even during extreme peak demand," Mallosh said.

Broader use

Beyond customer support, Floward is extending the same system to other customer interactions tied to sales. One example is an e-invitations feature that uses the platform for approval workflows, recipient notifications and gift prompts.

This points to a wider shift in how online retailers use messaging systems. What began as a customer support channel is increasingly being tied to order management, fulfilment and follow-on commercial activity within a single conversation thread.

Infobip, which provides cloud communications software, said Floward's use case reflected the type of workload for which multi-agent AI systems are most useful: high-volume, repetitive requests with clear handover points to human staff. It argued that the main benefit came from reducing fragmentation in customer journeys rather than simply automating single tasks.

"Floward's peak-season challenges are a textbook example of where agentic AI delivers the most value," said Emir Kalem, Head of Customer Success EMEA at Infobip.

"By orchestrating AI agents, WhatsApp Flows and live agent escalation on a single platform, we helped Floward turn fragmented customer journeys into a seamless experience that handles extreme volume spikes while improving end-user service quality and delivery performance," Kalem said.