Conversational Technologies
4 min read

Raising $20M to Advance Responsible AI-Powered Communications for Enterprise

Israel Krush CEO & Co-Founder, Hyro
Raising $20M to Advance Responsible AI-Powered Communications for Enterprise

Three Hypotheses

When Rom Cohen and I founded Hyro in 2018, we did so under three core hypotheses:


1. Conversational AI interfaces (chatbots, voice assistants, etc.) are the future of human-computer interaction (HCI).


2. The intent-based model of building conversational AI interfaces that was (and still is) widely adopted is unscalable, inefficient, and costly.


3. Deploying and maintaining conversational AI interfaces is prohibitively time and resource-consuming, especially for large enterprises with extensive data spread across multiple sources.


Five years later, it’s obvious that hypotheses 1 and 2 still hold true, and while some might argue that deploying conversational AI interfaces is now much easier thanks to the recent advancements in Large Language Models (LLMs) such as GPT-4, I believe that the barriers to entry for large enterprises wishing to implement conversational AI interfaces in a safe-guarded way are actually even higher than they were back in 2018.


So though it may seem as if natural language interfaces have become an off-the-self commodity, the fact remains that creating an enterprise-grade product that brings value and is worth the trust and dollars of major companies still involves rigorous engineering and development.


Or, in simpler words: It’s now easier than ever to produce an incredibly-impressive demo but just as hard as ever to build an average product.


At the end of the day, LLMs are not a magic wand, and while they simplify some aspects of general natural language understanding, they present many new challenges and lack three key components that would make them safe and viable for use by enterprises as standalone solutions:

1. Explainability

LLMs are essentially enormous ‘black boxes’; you know the inputs, you see the outputs, but there’s no telling what happened in between. Fixing errors starts with knowing how to find the problem, and with LLMs, there’s no way to investigate the root cause of wrong and misleading conversations.

2. Control

For enterprises, understanding the reasoning behind a conversational interface’s utterance is just one piece of the puzzle; the other piece is having complete control over the why, what, when, where, and how of that utterance. Multi-billion dollar companies demand conversational interfaces that follow strict business logic and feed off approved internal data sources. Lack of control opens up far too many loopholes for potential wide-scale customer friction, a non-starter for most organizations.

3. Compliance

Meeting the standards for HIPAA, GDPR, and SOC2 compliance as a conversational AI company is challenging enough as it is today, and by all indications, it’s going to get increasingly harder in the near future. As with every new wave of technology taking the world by storm, we can expect it to be followed by an even greater wave of regulation. Most newcomers we see flooding the market today will struggle to meet the strict demands of regulatory and governmental bodies. Only platforms with pre-built flexible safeguards in place to keep pace with ever-changing policy will be able to remain competitive. Compliance must be an inherent pillar of any legitimate conversational AI solution, not an afterthought.

An Adaptive Approach to Natural Language Interfaces

At Hyro (and I write this fully aware that hindsight is 20/20), we employed a wholly unique approach to natural language interfaces from day one. While many of our competitors and peers chose to tackle this challenge by building elaborate data pipelines, recruiting armies of data scientists, and constructing their own in-house natural language architectures (that are now being supplanted by an influx of new LLMS), we chose a much more adaptive approach.


We developed a plug & play platform that enables the orchestration of different components in a controlled, explainable, and flexible way. So when GPT-4 came out, we didn’t need to recalibrate our route. We treated it as a valuable add-on, not a threat. And today, we are using GPT-4 to unlock new use cases for our solution, recently releasing Spot™, a GPT-powered assistant capable of instantly generating answers to FAQs and search queries using customized enterprise data, all while maintaining explainability, compliance, and control. 


We understand that in this new era of AI, the x-factor will be the solutions provider that most responsibly and effectively blends key components of conversational AI, including emerging technologies such as Large Language Models, and we’re best positioned to accomplish that here at Hyro.

Building a Responsible Future

I have no doubt in my mind that the breakthroughs we’re currently witnessing in AI will fundamentally change how humans interact with computers and that the pace of advancements will further defy our expectations. At a time when innovation far outpaces legislation, it’s never been more important to advance responsible solutions that can be deployed by heavily-regulated industries such as healthcare. 


With this latest round of funding, Hyro will carry on to meet the healthcare sector’s critical demand for conversational AI and automation, enabling large organizations to streamline access to digital services for their patients and staff while drastically improving operational efficiencies. 


With 47% of the current US-based workforce in healthcare planning to exit the industry by 2025, we’re in a pole position to capitalize on soaring demand for AI and natural language understanding (NLU) solutions. As digital expectations for health services keep rising, health systems will need to compete for patient acquisition and satisfaction with a depleted staff, leaving Hyro’s AI-powered communications as a natural bridge for that growing labor gap.


As we look further into the future and with the backing of new investors, Macquarie Capital, Liberty Mutual Strategic Ventures, Black Opal Ventures, and K20, as well as existing investors Hanaco Ventures, Spero Ventures, and Mindset Ventures, we intend to expand our footprint to other regulated industries such as insurance and banking. We’ll recruit top talent across all departments to help us continue building our no-code platform for AI-powered call center, web, and mobile solutions while expanding strategic partnerships, integrations, and use cases.

Thank you to all who’ve supported our journey so far. 


Here’s to what’s next,


About the author
Israel Krush CEO & Co-Founder, Hyro

Israel is Hyro’s CEO & Co-Founder. Starting as a software engineer at Intel, he steadily progressed to leading engineering and product teams at various high-profile startups, including Zeekit, a computer vision company acquired by Walmart. Israel’s biggest love (following his wife and three children) is excellent coffee, which serves as the jet fuel for his bigger-than-life ambitions.