holy grail

Agent Efficiency: The Holy Grail of Contact Center Operations

The contact center landscape in 2024 is a foggy mix of aging IVR vendors desperately trying to hold onto their customers, consultants selling even more CX solutions, and smaller companies trying to barge in with point solutions, which I call CX “snake oil.” Sharp elbows, tall claims, and “AI washing” dominate the conversations.

A loyal customer asked me last year, “Your technology is automating calls, and my call volume for agents has dropped. But I still have agents; how can your generative AI technology help my agents holistically? You should take a fresh look.”

Since Avaamo had no legacy IVR product to protect, this kick-started some soul-searching. Did we want to wade into this foggy soup with yet another product? How would we differentiate with all the “AI washing” going on in the market? We talked to scores of call center managers and agents, discussing their journey, analyzing their in-call and post-call actions and workloads. Sound familiar?

contactcenter yesterday today

A Day in the Life of a Contact Center

The Agent: Addressing irate customers, adhering to tight work schedules, and adapting to erratic work hours.

The Call Center Manager: “We don’t have enough agents online.” Indeed, a veritable multi-billion dollar industry of vendors, consultants, and systems integrators push and promote this narrative, adding “micro features” to over-analyze and measure.

The Problem

“Agent assist” is a widely used term in contact centers, where vendors describe a feature that understands call intent and serves relevant information from a knowledge base. This typically includes suggested responses or articles. Some legacy vendors use basic search or a combination of manual scripting and machine learning, masquerading it as AI-powered Agent Assist. To us, the problem seemed centered around the sheer workload – the searches, the pop-ups, the notes, the data entry – trying to extract or enter data into antiquated systems while servicing an irate customer.

Avaamo’s Approach: Fixing the Problem – and Then Some

The product is the Agent experience. We designed a product that eliminated the searches, clicks, and notes by trying to anticipate agents’ needs in real-time and thoughtfully provide assistance using generative AI every step of the way. Our design automated more than 50% of an agent’s activities, such as searching, typing, and responding. We separated our design into four broad areas:

1. Dynamically Recommend Responses in Real-Time

Standard agent assist offers basic messaging and search support. While basic agent assist relies on keywords and predefined rules, generic responses are typically scripted and can address simple needs. However, they do not service intricate customer issues, particularly those that require personalized recommendations concerning returns, cancellations, or service interruptions. Combining our real-time transcription and LLaMB (generative AI) framework, our software excels in generating real-time recommendations “in the moment” for agents as the conversation evolves. They involve nuances like sentiment analysis, previous interactions, and specific customer profiles, which is precisely where agents need support.

2. Ingesting and Auto-Updating Content

Businesses are not fixed in time; things are constantly changing: new offerings, promotions, and policies. Manually updating knowledge bases and policy documents for agent use demands substantial and ongoing investment. It’s no wonder these systems frequently fall behind. Avaamo Agent Assist allows for the automatic update of multiple knowledge bases and policies directly from their source repositories, without intervention. This ensures that recommendations and responses are consistently current, eliminating any delays in agent support associated with reprogramming outdated rules-based approaches.

3. Automating Multiple Workflows and Actions for Agents

This work typically involves high-volume repetitive tasks, including collecting customer information, manually updating a CRM or other systems, setting appointments in separate systems, and recording “promises.” Avaamo’s Generative technology can automate time-consuming, multiple workflows for the agent. For instance, when an agent captures a change of address or initiates a cancellation procedure, the system automatically populates the CRM with accurate and error-free information.

4. Automating Call Summaries and Post-Call Activities

After a customer disconnects, there’s a significant amount of post-call work, capturing disposition notes and promises. These manual notes document the call’s details for analysis and promise management. Often, disposition notes are hurriedly written, prone to errors, and susceptible to incorrect categorization and inconsistent tagging, thus undermining their original purpose. Avaamo Agent Assist automatically generates summary notes, addressing the need for both human-readable and analytics-ready summaries.

Your Customers Want Results – Not Sympathy

We believe customers are fed up with boilerplate call scripts and fake expressions of empathy; they want straightforward outcomes with minimum fuss. If we can help the 8 million call center agents currently in the US and those around the world deliver on this need, we will have done something meaningful.

Check out the demo below or email us founders at avaamo with your feedback or comments. We’d love to here from you!


Ram Menon, CEO & Co-founder