The concept of customer experience has dominated discussions for years, centering on enhancing digital interactions between customers and brands. However, progress has been uneven. Sierra, the new venture from Bret Taylor and Clay Bavor, believes AI agents could represent the next significant technological advancement—similar to the pivotal roles played by websites and mobile apps. They envision these agents not just as valuable digital tools for companies, but as a means to truly fulfill the potential of digital customer experiences.
Regardless of skepticism, Taylor and Bavor perceive AI agents as a transformative tech category that can revolutionize customer-brand interactions and elevate overall satisfaction. “Our thesis is straightforward: we believe conversational AI will become the primary method through which consumers engage with brands, extending beyond current functionalities like customer service to encompass the entire customer journey,” Taylor shared.
In practical terms, this means customers can submit natural language questions or requests into a search box, and the AI agent will process these requests effectively, connecting to necessary transactional systems. Examples of these tasks include checking an order status or rescheduling a delivery.
The founders acknowledge that integrating with legacy systems can be complex. Nevertheless, most CIOs they’ve consulted have confirmed that APIs connecting to these older systems have been established, facilitating smoother communication for Sierra.
While Taylor and Bavor are excited about the potential of AI agents, they are also aware of significant challenges. “Introducing AI to customers undoubtedly enhances value, but it also increases risks, such as brand misrepresentation and hallucination—major hurdles that represent some of the toughest challenges in AI,” explained Taylor.
The hallucination issue, where AI models produce incorrect information due to gaps in knowledge, poses a serious threat to brand integrity. Although no solution has definitively resolved this challenge yet, Sierra is actively working to lessen its impact. The company’s technology relies on autonomous agents, meaning that responses from Sierra agents are generated using multiple models—sometimes as many as seven. One of these models, known as “the supervisor,” evaluates response quality, and questionable answers are sent back for reassessment. Taylor acknowledges that tackling hallucinations remains a pressing research challenge for the industry.
Moreover, handling automated customer data brings various regulatory and privacy considerations to the forefront. However, both founders assert that their agents are equipped to manage these complexities as well. Taylor contends that AI represents a departure from traditional software paradigms developed over the past three decades. He emphasizes the need for educational initiatives to help customers understand both the advantages and pitfalls of this new technology. “Our market approach includes risk mitigation along with educating customers on how this type of software operates,” he stated.
Despite the inherent risks, Taylor and Bavor view this period as a remarkable opportunity for innovation. “When technology landscapes shift, it creates a space for startups to experiment and take calculated risks,” Bavor noted.
According to Taylor, this wave of AI innovation is likely to spawn five to ten new independent enterprise software companies, mirroring the emergence of mobile and cloud solutions. “We’re in the early stages of defining the market for conversational AI. There’s currently no dominant player because this field is just taking shape,” he articulated.
Taylor, who also serves as board chair of OpenAI, insists that there is no competitive conflict between the two organizations. "We don’t perceive OpenAI as a competitor, and I will recuse myself in any situation that presents a conflict,” he affirmed.
The founders also advocate for a new pricing strategy tailored to the outcomes achieved rather than traditional subscription fees or usage-based rates. Their vision centers on clients paying solely for results delivered, promoting a shift toward outcome-based pricing models. “We believe outcome-based pricing is the future of software. With AI, we finally have technology that not only boosts productivity but also completes tasks effectively,” Taylor indicated, although the specifics are still being refined with early users.
Despite their expertise, industry analyst Brent Leary from CRM Essentials expresses skepticism about competing against established players like Salesforce. “Taylor is undoubtedly brilliant, but Salesforce possesses institutional knowledge, resources, and R&D investments that a startup simply lacks. Such large companies are rapidly adjusting their operations to harness AI opportunities,” Leary remarked.
Sierra may not have the same extensive resources as giants like Salesforce, but the combined credentials of Taylor and Bavor have garnered significant interest, leading to $110 million in financing. This includes an initial $25 million from Benchmark, followed by an $85 million second round from Sequoia. Such funding is substantial for an early-stage enterprise, particularly under the guidance of notable founders.
Ravi Gupta, a partner at Sequoia Capital leading the investment in Sierra, stated that the firm was captivated by the technology's potential after witnessing it firsthand. “Seeing it in action was truly impressive, and it really ignited our vision of what future customer interactions could look like,” he added, confirming it was an easy decision to invest.
Sierra recognizes a significant chance to redefine customer experience through AI, yet numerous challenges remain. If the founders can effectively navigate the complexities of providing free-form, AI-driven customer interactions while remaining competitive against established enterprises, they could establish a successful startup. Yet, as with all things AI-related, they must continually demonstrate consistent, scalable results to solidify their position in the market.
Ranking
