Opinion
A digital accountability lesson: The bitter truth companies must learn from the cloud outages before Black Friday
"Ahead of the November shopping holidays, my recommendation to executives is this: Don't settle for asking, 'How smart is my AI bot, or how can I advance further with AI?' Also ask, 'What happens to my customer service center when the internet breaks or the cloud fails?',” writes Gidi Adlersberg, Business Line Manager at AudioCodes.
Just recently, we received yet another painful reminder that the internet, for all its power and grandeur, is fundamentally a network of data pipelines. When a central pipeline, like that of Cloudflare clogs or leaks, half the world finds itself paralyzed. This is not an isolated incident; in the past month alone, we witnessed significant outages among giants like AWS and Azure. For the average consumer, these disruptions might have amounted to difficulty browsing a favorite news site or an error message on a banking app. But for CIOs and VPs of Customer Service, the ensuing silence was far louder and more alarming.
In recent years, we have seen a remarkable arms race in the field of Customer Experience (CX), with companies investing hundreds of thousands, if not millions of dollars in services and communications. Today, almost all of our core digital services - from company websites and applications to Customer Relationship Management (CRM) systems - and the sophisticated Voice AI solutions that power them (based on Large Language Models and STT engines) reside in the public cloud. The moment a major cloud provider goes down, this entire digital network can be silenced. The brilliant bot, which serves millions of customers, suddenly becomes helpless, websites crash, and the customer is left on a silent line or facing an error page.
This recent cloud failure serves as a crucial wake-up call, arriving just before we enter the retail world’s "money time": November, Black Friday, and Cyber Monday. During this period, traffic loads on customer service centers surge by hundreds of percent. Customers call to check shipping statuses, process returns, or make last-minute purchases.
Imagine a scenario where, in the middle of the sales frenzy, a cloud outage disconnects your contact center. The damage is not only immediate revenue loss but a devastating blow to brand trust. A customer who encounters a disconnect or a non-functioning customer experience under pressure is a customer who churns. They are here to buy a product; they do not care which cloud system their payment is processed on, and they will likely just move to the next available competitor.
So, what should be done? Is the solution to revert to the rigid, legacy technologies of the 1990s? Certainly not. But in our enthusiasm for cloud, Generative AI and the revolution in conversational commerce, we have neglected a vital, age-old concept: Resilience and Business Continuity.
Globally, many organizations are still designing their core architecture, including their contact centers and Voice AI solutions, somewhat naively, assuming the cloud will be available 100% of the time. The lesson from these recent failures must be the adoption of smarter, hybrid models. Just as we install a physical generator in case of a power outage, core digital systems must have an active and advanced safety net. To safely navigate November (and business generally), organizations must ensure that even if the central infrastructure collapses, the service does not go silent.
The solution lies in implementing true Redundancy: Modern cloud and AI systems must be designed to automatically fall back to a resilient and available service in a fraction of a second. This includes the critical capability to switch from a cloud service to a local (on-premise or Edge) service that can function as a backup in times of crisis, should the cloud go down.
Furthermore, it is essential to ensure a persistent Voice Safety Net: The customer service phone line itself must remain open and available. It needs the capability to detect a failure in the contact center or AI bots immediately, and execute an immediate tactical fallback - transferring the call to a human agent through a simpler, local contact center backup system.
In a dynamic era where everything moves fast and competition is fierce, every single infrastructure failure critically disrupts both the customer and the service organization. Technology continues to lead in developing breakthrough customer service solutions - from bots that understand sarcasm and speak natural language to those that solve complex problems. But these systems only function when the stage is stable. We must not let innovation blind us to the fundamentals: first and foremost, we must ensure availability.
Ahead of the November shopping holidays, my recommendation to executives is this: Don't settle for asking, "How smart is my AI bot, or how can I advance further with AI?" Also ask, "What happens to my customer service center when the internet breaks or the cloud fails?”. The answer to that question will determine whether you end the quarter with a rising sales graph or a massive reputational crisis that will take many months to fix. Innovation is already here; it just needs to be more accountable, more resilient, and more grounded - even when it is in the cloud.
Gidi Adlersberg is Business Line Manager at AudioCodes, specializing in advanced Voice AI solutions for enterprises.
