AI is everywhere, but people are still trying to pin down tangible, lasting use cases. Leveraging conversational AI to create what I call a “25-hour receptionist” – i.e., AI that goes above and beyond human capacity to solve critical needs in specific vertical use cases, like the ones I saw as CEO and co-founding investor of Presto Automation – has the potential to be one of the most immediately actionable and highest-value AI applications. While firms like a16z have recently validated the opportunity of these applications, our experience with Presto and other companies has shown me it’s an area full of nuance around the Human-AI Waltz, the ROI to customers, and verticalization.
The simplest AI-related question one can ask is: can this tool replace labor to perform a repetitive, typically customer-facing task? For conversational AI companies, this usually means automating a “front desk” or “first interaction” role that humans typically complete. That automation enables a customer to have a “24-hour receptionist” – an on-call conversational agent who never misses a call or an order using a combination of pure AI and, for now, a human in the loop. That could mean scheduling appoints/automating patient intake in healthcare organizations, automating responses to customer calls in retail, or automating order-taking in restaurant catering or drive-thrus, as our company Presto does. In this latter example, the Presto Voice AI system can take a customer’s order in the drive-thru in place of a restaurant employee. At a time when California’s fast food minimum wage has increased to $20/hour, there is a powerful need for this kind of tool.
Automating first-interaction tasks is already of value, as it’s nearly impossible for a set of humans to work ‘24-hour’ shifts consistently, affordably, and in a way that facilitates a positive customer experience. Leveraging this application of AI requires a carefully honed waltz between technology and humans to credibly replicate an intelligent human conversation that achieves both accuracy and efficiency.
But, while it’s certainly challenging for human operations to ensure 24-hour responsiveness in a consistent manner, it’s not impossible. AI that creates a “24-hour receptionist,” then, is very tangible in the value it creates, but ultimately limited in dollar value, especially as it becomes commoditized over time. While alluring, I don’t think this straight labor replacement value proposition is the “killer app” of conversational AI.
The 25-Hour Receptionist and the AI Premium
The real, sustainable and nearly unlimited power of these conversational AI agents comes from something more: what I call the “25-hour receptionist,” or the ability for AI to do things a traditional human in the role could not. The “AI premium,” that extra hour, is much more possible with vertical applications and is what creates sustainable defensibility for these companies – as the technology gets better, it effectively transforms the receptionist into a concierge. That extra hour creates more value than the other 24 hours combined. I am looking for companies building verticalized 25-hour receptionists and dreaming about building 25-hour concierges.
In practice, a “25-hour receptionist” can look like AI that understands and deciphers customers’ needs as a personal shopper in retail, that triages patient symptoms as a nurse rather than just a receptionist in healthcare, or that makes real-time, targeted upsells at the drive-thru.
The perfect 25-hour receptionists find ways to drive additional value for end customers and thus creates more value for employers as well. They have endless information available to personalize and customize the customer experience in real-time without putting the customer “on hold” and have emotional intelligence in a way high-performing humans do. AI that can go above and beyond by intaking more information, better understanding customer needs, and providing personalized suggestions can leave the customer feeling particularly well-served. The individual elements of this unified experience are all now technically feasible and in use at successful companies, but the 25-hour receptionist pulls them together in a real-time, cutting-edge application of conversational AI.
These 25-hour receptionists do NOT try to go fully upstream and capture all the value; thus, they are not the doctor, nor the store owner, nor the chef. Contrary to what some believe, I don’t think the technology or the customer appetite is there to replace these higher value interactions anytime soon. The beauty of LLMs, though, is that they have unlocked a significant first part of the value chain.
Not every vertical conversational AI company will be successful in envisioning and executing on this 25-hour receptionist, let alone the concierge. Some will become mired in crafting the perfect 24-hour receptionist and others will fail to uncover use cases for that AI premium. Horizontal AI companies will find it much more challenging to unlock this premium, as such applications are usually highly vertical-specific, requiring vertical integrations, data sets, and customer insights.
In nearly every company in nearly every vertical, there is an opportunity first to automate customer interaction, then to ascend from 24- to 25-hour responsiveness and value-creation. Given my experience working on this at Presto, I’m acutely focused on the vision of the founding team and am curious how they’re building toward a dynamically automated future – one that will for the foreseeable future remain a waltz between humans and AI.
If this post resonates with you, reach out!