In the fast-evolving world of AI, a fresh contender is emerging that could challenge the dominance of neural networks. This alternative is known as symbolic AI—an approach that builds on predefined rules tailored to specific tasks, such as rewriting text or automating workflows, to solve more complex problems. While neural networks like OpenAI’s ChatGPT have dominated the AI landscape, symbolic AI is showing promising potential, especially in domains where conventional AI systems tend to falter.
Historically, symbolic AI was criticized for being inefficient and lacking the scalability of neural networks. But recent breakthroughs in scaling symbolic architectures have ignited a surge of interest from startups seeking to leverage this technology. Companies like Orby and TekTonic are using symbolic AI to build enterprise automation tools, while others like Symbolica and Unlikely AI, founded by Alexa co-creator William Tunstall-Pedoe, are exploring new applications. One of the latest entrants into the field is Augmented Intelligence, a startup that has recently come out of stealth mode with $44 million in funding from a roster of high-profile investors, including former IBM President Jim Whitehurst.
Symbolic AI: The Secret Behind Augmented Intelligence
Augmented Intelligence focuses on building conversational AI that aims to be more predictable and action-oriented—what its CEO, Ohad Elhelo, calls “agentic.” In other words, it doesn’t just respond to questions with generic information; it takes action on behalf of the user. For example, rather than just providing a list of available flights, the AI can compare prices and book a flight for you—all with minimal setup.
At this point, you might be wondering: Can’t ChatGPT or other conversational AI already do that? Yes, they can. But according to Elhelo, the difference lies in the efficiency and ease of integration. He claims that platforms like ChatGPT require more manual integration and setup to deliver those services, while Augmented Intelligence’s technology works seamlessly with existing tools, requiring less intervention from users or businesses.
“There’s a big difference between chatbots like ChatGPT, whose primary goal is to engage in conversation, and conversational agents that actually take action,” Elhelo explained in an interview with TechCrunch. “Once you connect the AI to external tools—either to retrieve information or to execute a task—most models struggle. The quality of intelligence drops dramatically because they stop relying on their training data.”
From Delegate to Augmented Intelligence: A Journey of Pivot and Reinvention
Elhelo’s AI journey began in 2017 when he co-founded the company with Ori Cohen. Back then, the startup was called Delegate (and previously, Stuff) and had a very different—and more controversial—mission. Delegate was designed as an AI-powered app that allowed users to outsource tasks to low-paid gig workers. Elhelo once pitched it as a way for busy professionals to save time: “If an hour of a business executive’s time is worth $50, why spend it on something like customer service when there’s someone in India or the Philippines willing to do it for much less?”
However, Delegate’s reputation took a hit after a flood of negative Glassdoor reviews from workers, who complained about a buggy platform, lack of training, and an opaque incentive structure. Faced with these challenges, Elhelo and Cohen made the strategic decision to pivot toward AI-driven solutions, ultimately evolving into what is now Augmented Intelligence.
What Sets Augmented Intelligence Apart?
Today, Augmented Intelligence develops AI that can power chatbots capable of answering a wide range of customer inquiries—such as “Do you price match on this product?”—and integrate seamlessly into a company’s existing workflows and APIs. Unlike many neural network-driven systems, Augmented Intelligence’s AI draws on symbolic AI techniques to access and use external tools, giving it a distinct edge in complex, multi-step processes.
For example, while OpenAI and other companies have enabled AI to use tools, Elhelo argues that Augmented Intelligence is simply better at doing so. Moreover, the AI is designed to be more explainable. Elhelo asserts that companies using the platform can access detailed logs of how and why the AI made specific decisions, allowing for fine-tuning and improvements over time. He also emphasizes that Augmented Intelligence’s AI doesn’t train on a company’s data, only using the information and tools that the company explicitly allows.
“Augmented Intelligence doesn’t require additional training on company-specific information,” Elhelo said. “Instead, it follows the rules and instructions provided by the business, ensuring compliance and reducing the risk of data exposure.”
This feature could make Augmented Intelligence especially attractive to businesses wary of sharing sensitive data with third-party AI providers. Companies like Apple have reportedly banned the use of some AI tools, such as OpenAI’s, out of concerns over confidential data leakage. Augmented Intelligence’s ability to bypass the need for proprietary data training might give it an edge with these more security-conscious clients.
The Road Ahead: Can Augmented Intelligence Compete with the AI Titans?
While Elhelo has made bold claims about the capabilities of his company’s AI—such as its supposed ability to eliminate hallucinations (a common issue in neural network-driven AI models)—there’s no denying that Augmented Intelligence has momentum. The 40-person startup recently formed a strategic partnership with Google Cloud, signaling growing interest from major industry players.
Though Elhelo has declined to share revenue figures, he did reveal that the company’s most recent funding round, led by New Era Capital Partners, raised $10 million and valued the company at $350 million. That’s an impressive valuation for a relatively new AI vendor, especially one that wasn’t founded by a household name in the AI industry.
As for the future, Elhelo is confident that Augmented Intelligence’s neuro-symbolic architecture, which combines the pattern-recognition strengths of neural networks with the structured logic of symbolic AI, will continue to unlock new possibilities for businesses.
“Traditional models excel at pattern recognition and language generation,” he said, “but they fall short when it comes to performing actions, making decisions, or interacting with tools. Our approach solves those problems and gives companies a powerful way to automate complex tasks.”
With a compelling value proposition and backing from some of the tech world’s most influential figures, Augmented Intelligence may indeed be poised to carve out a space for itself in the increasingly crowded AI landscape. The company’s focus on symbolic AI, action-oriented conversational agents, and explainable models could offer businesses a fresh alternative to the neural network-dominated AI tools that have become so familiar today. Whether or not it can truly compete with industry titans remains to be seen—but one thing’s for sure: Augmented Intelligence is a company to watch.
