The $100 Billion Shift in Customer Service: How Ada is Leading with AI

The world of customer service is undergoing a major transformation, and Ada, a company founded in 2016, is playing a pivotal role. Valued at $1.2 billion, Ada has raised a total of $200 million and is revolutionizing how businesses interact with customers. Major brands like Verizon, YETI, Canva, and Square are among Ada's clients, all benefiting from their AI-native customer service automation platform. So, what sets Ada apart in this rapidly changing landscape? Let's break it down.

The Evolution of Ada’s AI Journey

Ada has always been an AI-first platform. From its inception, the company used custom Natural Language Processing (NLP) models to handle customer inquiries. But they quickly realized something was missing. While their platform could handle many customer queries, the quality of resolutions often fell short. As Mike Gozzo, Chief Product and Technology Officer, explains:

“We got really excited by OpenAI and what was happening in the industry. In 2022, we decided to go all in and rebuild the product using the reasoning capabilities of large language models (LLMs).”

Redefining Customer Service with a New Metric: Resolution Rate

Traditionally, customer service performance was measured by a metric called containment rate. This metric looks at the percentage of customer inquiries handled entirely by a chatbot without needing human intervention. While a high containment rate might reduce operational expenses, it doesn’t guarantee a positive customer experience. As Gozzo puts it:

“Many solutions, including Ada, could easily deliver 80–100% containment rates, but if you actually opened up those conversation transcripts and read the experiences customers had, they were quite poor.”

To address this, Ada developed a new metric: resolution rate. Instead of focusing solely on containment, they prioritized how well customer conversations were actually being resolved. By using GPT-4 and Ada’s own historical data, they built a system capable of evaluating conversations based on how well they answered customer questions with relevant, accurate, and safe responses.

Battle-Testing Models with OpenAI

Once Ada had resolution rate as their guiding star, they knew they needed a powerful AI partner to keep raising the bar. They chose OpenAI. Ada used OpenAI's models before, but mostly for generating training data. However, with their new focus, Ada tested OpenAI’s capabilities more rigorously.

Ada has developed a synthetic test framework that simulates conversations with hundreds of thousands of customers to stress-test AI models on quality and scalability. As of now, OpenAI’s models remain the gold standard, especially in terms of quality of inference and low-latency performance. “No one has bested OpenAI on our evaluation set,” says Gozzo.

Ada also uses OpenAI’s fine-tuning capabilities to improve model accuracy, reducing hallucinations (incorrect responses) and ensuring that AI-generated answers are as reliable as possible.

Doubling Resolution Rates with OpenAI

The real game-changer for Ada has been the impact on customer resolutions. Before the shift to OpenAI, Ada’s product had a containment rate of 70%, but only a 30% resolution rate. With the new system in place, containment rates are similar, but resolution rates have skyrocketed—often reaching up to 60%, with top-performing clients seeing resolution rates of more than 80%.

For Ada’s customers, this improvement translates into significant benefits—lower operational costs, increased customer satisfaction, better retention, and higher rates of new signups.

The Road to 100% Resolution

What’s next for Ada? They’re aiming for nothing less than a 100% resolution rate, a goal that once seemed impossible. “Now, 100% resolution is a when, not an if,” Gozzo states confidently.

And it’s not just Ada that’s feeling optimistic. Enterprises are becoming more savvy about AI and eager to incorporate automated resolution systems into their operations. Ada’s AI Agents are increasingly seen as essential tools that businesses can train and manage like human agents.

Ada is looking forward to expanding the transparency and control it offers to its customers in the coming year, making AI-powered customer service more accessible and effective than ever before.

The Implications of Ada’s AI-Driven Shift

For those deeply invested in the AI space, Ada's advancements signify a pivotal moment. Here's why this matters:

1. The Evolution of AI in Customer Service: Ada's shift from traditional NLP models to advanced LLMs like GPT-4 illustrates how rapidly AI is evolving. It’s no longer just about handling queries, but doing so with quality, accuracy, and at scale. This sets a new standard in customer service and will likely push other companies in the industry to follow suit, which means AI development will intensify to meet these higher expectations.

2. A Shift in Workforce Dynamics: As Ada improves resolution rates with its AI systems, fewer human agents are needed to manage routine customer inquiries. This trend will significantly alter the customer service workforce. Companies will be more reliant on AI, leading to potential job shifts or reductions. While AI creates opportunities in development, training, and maintenance, it also means that lower-skilled customer service jobs may become less prominent.

3. The Role of AI Enthusiasts and Experts: For AI professionals, the rising demand for fine-tuning models and reducing hallucinations means there’s growing importance for specialization. Companies will need experts who can ensure AI performance remains accurate, safe, and efficient, which opens doors for AI experts in fields such as AI ethics, model optimization, and fine-tuning.

4. New Metrics and Industry Standards: Ada’s introduction of a new customer service metric, resolution rate, highlights how AI is reshaping industry standards. The containment rate, once a widely accepted measure, now seems outdated. As AI solutions continue to grow more sophisticated, other industries beyond customer service will need to rethink how success is measured when AI is at the core of their operations.

5. AI Impact on Customer Experience: From a consumer perspective, the improved resolution rates powered by AI like Ada’s mean smoother, faster, and more satisfying interactions with businesses. AI enthusiasts should note that these customer service innovations will raise expectations for AI-driven experiences across all industries, potentially accelerating AI adoption everywhere—from healthcare to retail.

Final Thoughts: How Will People Be Impacted?

The positive side of Ada's innovation is the potential for more efficient and high-quality customer service, where AI handles tasks faster and more accurately. On the flip side, people in customer service roles may feel the pressure of automation displacing traditional roles. However, those willing to adapt and reskill can find new opportunities in managing, fine-tuning, and optimizing these AI systems.

In summary, Ada’s approach is not just a technical upgrade—it's reshaping the future of work, creating new industry benchmarks, and redefining customer expectations in an AI-driven world. For AI enthusiasts, this is both exciting and challenging, as it presents new opportunities and highlights the need for continuous learning and adaptation.

If you want more updates related to AI, subscribe to our Newsletter