Montreal startup HumanFirst raises $5M to transform organizations’ conversational data into no-code AI

Imagine you run a customer service desk at a company. You quickly notice that your employees spend a good chunk of their day answering the same questions from different customers. Wouldn’t it be nice to automate those responses and let AI answer them, freeing up your human employees’ time to handle trickier questions and situations?

But unless you have a programming background, creating such an automated system might seem out of reach. At least, it was until HumanFirst burst onto the scene.

A Montreal-based tech startup, HumanFirst offers a platform for creating new enterprise applications and processes based on a company’s own conversational data — that is, records of conversations between customers and support staff — using a no-code approach, meaning you don’t need to have an advanced knowledge of programming or computer science to use it. “The fastest way to build custom AI you can trust,” is how the HumanFirst website puts it.

The company today announced $5 million in seed funding from three smaller “safe-note” financing rounds, led by Panache Ventures, and joined by Inovia, Real Ventures, BoxOne Ventures, and angels including Lookout founder Kevin Mahaffey.

The new capital injection is set to fuel HumanFirst’s growth and customer acquisition. HumanFirst is targeting a doubling of its current small team size of 15 and a tripling of its current customer base, around 33, using the funding.

HumanFirst’s approach to no-code

The no-code movement, according to Gregory Whiteside, cofounder and CEO of HumanFirst, is fundamentally reshaping the software industry. It allows individuals lacking formal programming expertise to develop their own applications.

“HumanFirst stands as a crucial bridge to this new era, unlocking the value of conversational data, which is typically hard to decipher and utilize,” Whiteside in a press release.

Meanwhile, as interest in generative AI explodes, businesses are flocking to the tech, but many are still hesitant and cautious about deploying it in a meaningful way, in part due to perceived lack of in-house AI knowledge.

>>Follow VentureBeat’s ongoing generative AI coverage<<

Organizations by and large have lots of data that could be leveraged to build AI apps that help them — but they don’t know how to go about it safely and securely. That’s where HumanFirst comes in.

It allows businesses that engage in a high volume of customer interactions to automate those intelligently, using simple but powerful software workflows. In essence, it creates AI trainers to help surface and identify trends in large datasets, then test how the design of the data actually works in real time with natural language understanding (NLU) and large language model (LLM) inputs.

The AI responses are all based on a business’s conversational data, ensuring they are customized to match the recurring issues that come up between the business and its customers. The data that HumanFirst analyzes to allow customer businesses to build their own apps include call transcripts, emails, customer feedback and support tickets.

“We also integrate directly with different conversational databases like Rasa, Dialgflow and more,” Whiteside told VentureBeat in an email.

Panache Ventures’ Partner, Scott Loong, commended HumanFirst tech that allows businesses to extract and operationalize customer and business insights from large conversational datasets. “HumanFirst’s data-centric approach to simplifying and improving AI models for both technical and non-technical users is unprecedented,” Loong said in a statement.

>>Don’t miss our special issue: The Future of the data center: Handling greater and greater demands.<<

HumanFirst deploys “AI trainers” within its customer organizations, which smartly cluster and bucket information, allowing employees at the customer organization to observe and discover trends and insights. These trainers merge employees’ human-context and perspective with mass labeling and indexing of the growing dataset quickly and effectively, a task traditionally done on spreadsheets.

Early customer win

HumanFirst’s solutions are already streamlining business operations for various customers across industries.

“HumanFirst has a great list of enterprise customers including banks, call centers, and companies that trust the product and have built AI/automations that save time, money and uncover richer insights,” Whiteside told VentureBeat via email.

Intelcom, Canada’s last-mile delivery leader, is among the initial batch of clients currently using HumanFirst. Jean-Sébastien Joli, CEO of Intelcom, said that HumanFirst had already transformed his company’s call centers, allowing them to more rapidly identify customer issues and begin solving them without wasting customers’ or support staffs’ time trying to understand one another.

“HumanFirst has revolutionized our ability to understand data at scale. Their data-driven insights are integral in crafting superior customer and product experiences. The incorporation of HumanFirst’s automation recommendations in our call center is projected to reduce repetitive interactions by up to 50% during the initial phase,” he said in a press release.

Whiteside said he believes a human will always remain in the loop, but that HumanFirst can democratize access to building AI and make dealing with large, scaling datasets feasible for more organizations. At the same time, he did not sugarcoat the point that his customers might use his technology to reduce their human staff.

“If we can unlock the potential of richer insights and better collaboration across teams — we think that companies may reduce the number of people in some areas of the business, but also, can empower talented people to observe, launch and build impactful products and experiences for their business in ways they couldn’t dream of before,” Whiteside wrote to VentureBeat.