Beyond Q&A: Teresa Torres Highlights AI’s Gap in Complex Customer Support

In a recent LinkedIn post, Teresa Torres delves into the limitations of current AI support agents, particularly when faced with complex customer service scenarios that extend beyond simple question-and-answer interactions. Torres points out that while many AI systems excel at providing information, they often falter when customers require multi-step, procedural assistance.

The Challenge of Complex Support Scenarios

Torres highlights a critical gap in the capabilities of most AI support tools. She explains that the true challenge lies not in answering basic questions, but in handling intricate customer needs that involve multiple actions and verifications. As Teresa Torres notes:

“Most AI support agents can answer questions—but what happens when a customer needs their card frozen, their address verified, and a replacement ordered? That’s where most AI falls short.”

This observation underscores a significant area where AI currently struggles to match human agents’ ability to navigate and execute a series of interconnected tasks. The implication is that for a substantial portion of customer support work, AI remains insufficient.

Gradient Labs’ Approach to Procedural AI

The post then introduces Gradient Labs as a company that has developed a solution to address this very limitation. Teresa Torres explains their innovative approach, which centers around the concept of “procedures.” According to Torres, these procedures are designed to empower non-technical subject matter experts to define and manage complex workflows.

Empowering Subject Matter Experts with ‘Procedures’

Torres elaborates on how Gradient Labs enables a more sophisticated level of AI interaction. She describes their system as:

“natural language instructions that look like Notion docs. Non-technical subject matter experts can write step-by-step workflows, complete with tool calls, without touching code.”

This method, as outlined by Torres, allows for the creation of detailed, step-by-step instructions that AI can follow. By using familiar interfaces akin to document collaboration tools, Gradient Labs aims to make AI implementation more accessible and effective for handling the nuanced demands of customer support.

Unlocking the ‘90% of Support Work’

The core argument presented by Teresa Torres is that this procedural approach is key to unlocking the vast majority of customer support tasks that are currently beyond the reach of basic AI. She posits that by enabling the creation of these natural language workflows, companies can significantly enhance their AI’s capacity.

In Teresa Torres’s view, this innovation is crucial for businesses looking to leverage AI more comprehensively. It moves beyond simple chatbots to systems capable of executing complex sequences of actions, thereby addressing a much larger segment of customer service needs. The post concludes by pointing readers to further resources, including links to a podcast episode discussing these topics on Spotify, Apple Podcasts, and YouTube.