In an era where artificial intelligence is starting to disrupt many industries and markets, not every company says they use machine learning and AI in the same way. Companies that are able to truly understand and use these new tools in ways that deeply support their products or services will stand out from companies that lack the proper understanding and simply wrap the term “AI” around more traditional tools. A big part of this emerging differentiator is defining and understanding the problem space so that these advanced tools can be applied in meaningful ways.
These are the kinds of issues that Zayd Ali, the 21-year-old founder of Valley, spends a lot of time on. Valley is a fast-growing startup in the sales development industry focused on automating the $62 billion-a-year appointment booking market in a way that simply wasn’t possible even a year ago. The company develops and uses proprietary generative AI models and algorithms to reinvent business-to-business sales interactions.
This isn’t Ali’s first rodeo; Despite his young age, he has previously founded, grown and exited two other companies in the sales industry. Most recently, he exited a company he founded called Advisor Appointments through a private equity sale. Ali uses what he learned there as a springboard for his current venture.
How Valley uses generative AI
Zayd Ali’s path to innovation in sales technology began as a non-technical founder, a title that often comes with its own set of trials, especially in the tech-dominated landscape of Silicon Valley. However, Ali’s previous successes armed him with an arsenal of industry-specific knowledge. This expertise became the cornerstone of Valley’s strategy to disrupt the sales automation market. He was able to identify and carefully structure an industry-specific open problem that machine learning and AI cannot solve on their own.
At its core, Valley uses state-of-the-art generative AI and large language models to automate and streamline the tedious processes involved in identifying, developing and capturing business-to-business agreements. The idea was to build a system that could simulate sales development representatives in different industries, a very technically challenging goal that required pushing the boundaries of existing AI technology.
Generative AI refers to algorithms that can learn from a data set and generate new, similar data points. Depending on how it is trained and used, it can be an effective tool for machine inference, connecting the dots between data points in a meaningful way. Large language models, such as ChatGPT, are a subset of generative AI, aimed at understanding and generating human language and dialogue interactions from massive data sets.
In the context of Valley, this meant creating responses and interactions that a human would perform when making appointments. A very non-trivial task, but one that, given the right industry and domain expertise to provide context, large language models could be particularly well suited to tackle.
The ‘quality over quantity’ approach to building an industry-specific generative AI tool
Valley’s approach was to develop its own model: the Valley Reinforced Learning from Sales Feedback model. Unlike the typical “bigger is better” approach to data modeling, Valley is committed to specificity and accuracy. The model is not trained on all possible sales interactions, but on the most effective ones, honing its ability to emulate successful sales reps in a way that continues to learn and improve through iterative feedback loops.
In fact, the Valley team taps into a very active area of research when it comes to large language models and generative AI. One of the biggest limitations of extremely large models, such as those underlying ChatGPT, is the sheer size of the networks and billions of parameters that must be trained in the model. Training the network alone costs millions of dollars, let alone the cost of maintaining it and running queries on it. Furthermore, the physical resources required to train and operate these massive models, things like electricity and cooling, pose a challenge to the continued scaling such models can achieve.
As such, much research, both in academia and industry, is focused on developing smaller models and networks that can achieve comparable or better performance compared to larger, more general models as the quality and specificity of the data increases. This is a case where ‘good enough’ really is good enough.
Valley’s approach to achieving market and technological differentiation has been to focus on a strategy of data curation – the idea that with each interaction the system becomes smarter, more efficient and therefore more valuable to the individual user. This creates high switching costs for users because the more they use Valley’s system, the better it understands and meets their specific needs, discouraging them from turning to competitors.
Benefit from industry data and experience
Developing such a system required overcoming significant hurdles, from algorithm design to data collection and processing.
The team, consisting of individuals from Samsung AI, Columbia University, Salesforce, Yext, and more, began by building on the application layer of existing generative AI frameworks. This layer is responsible for adapting the model’s broad capabilities to specific tasks – in this case, making appointments. Creating the workflow involved setting up sequences of interactions that would most likely result in a successful appointment, just as a human representative would achieve through experience.
But unlike other startups trying to leverage generative AI, Valley didn’t rush to build the largest possible data set. Instead, they focused on assembling a high-quality data set, with each data point properly labeled and relevant to the appointment setting task. That data specificity meant their proprietary model could engage in high-fidelity interactions with promising leads, tailored to the nuances and needs of different industries.
They achieved this in part through the use of reinforcement learning, a type of machine learning in which an algorithm learns to make decisions by assessing outcomes as it learns. Valley’s model is continuously updated with feedback from sales results – a reinforcement learning strategy that ensures the AI’s responses are continually refined to be more effective in real-world sales scenarios.
Growing success by meeting the specific needs of industry challenges
The sales technology industry faces a range of challenges, from high customer acquisition costs to the need for scalable, predictable revenue. Valley is looking to address these head-on by significantly reducing the cost of acquiring new appointments and, by extension, the cost per conversion. This optimization offers companies new efficiencies and scalability that were not possible with traditional methods.
As Ali explained: “There are only three options for companies looking to make appointments with potential buyers – and none of them were optimal. One of these is founder-led prospecting, where if you have an early-stage company you can spend two to three hours a day connecting with potential buyers – a huge waste of time. The second option is to hire an appointment shipping agency for $3,000 to $4,000 per month – very expensive and many of them rarely do their job well. And the final option is that if you have the capital, you can build a full sales development team, which when scaled quickly becomes a seven-figure annual investment.
So the experiment with Valley was whether we could turn that $85,000 per year salary for a sales development representative or that $4,000 per month agency budget into a software cost of $400 per month. Build a product that can turn a cold stranger into a book sales meeting without any human intervention, and by doing so, dramatically change the customer acquisition calculus that our customers performed. And dramatically change the allocation of time on prospecting versus other critical areas of a start-up business.”
Ali declined to share revenue figures, but he stated that since starting their pilot program in March this year, they have grown 30% month-on-month and reached seven figures in signed letters of intent to expand existing customers.
Even as they continue to grow, their existing clients include Darwinian Ventures, Front.com and Masterworks. Valley recently announced raising a $2 million pre-seed round from investors including Antler, Jason Calacanis, Rough Draft Ventures, O’Shaughnessy Ventures, ID8 Investments, Transform VC and John Pleasants, the former CEO of Ticketmaster, Match and Evite.