As the realm of conversational AI continues to expand its influence, chatbots are experiencing a resurgence in relevance. Nonetheless, there are evident perils on the horizon. DPD, a prominent parcel delivery company, found itself compelled to deactivate a segment of its online support chatbot following an incident where the bot used inappropriate language towards a customer.
A user successfully manipulated the chatbot into expressing severe criticism of DPD and even persuaded it to articulate the critique in the form of a haiku.
This occurrence serves as a stark warning, particularly for industries subject to stringent regulations such as healthcare and insurance. Entrusting crucial interactions to a poorly designed AI chatbot in these sectors could potentially expose organizations to legal repercussions. OpenDialog, a UK-based startup founded by seasoned entrepreneurs, believes it has a solution by merging cutting-edge LLMs with a customizable platform tailored for regulated fields.
The company recently secured \(8 million in a Series A funding round led by Alboin VC, with additional contributions from Dowgate Capital and several prominent angel investors, bringing its total funding to \)13 million.
The demand for conversational AI is on a meteoric rise, projected to reach a staggering $38 billion globally by 2029. Nevertheless, regulated industries are still grappling with the complexities of Natural Language Understanding (NLU) and Large Language Models (LLMs). OpenDialog asserts that its specialized focus on healthcare and insurance positions it as an ideal partner for these sectors.
While an LLM may possess the ability to emulate human language and grasp context, industries governed by regulations necessitate stringent controls when adopting an AI-driven approach. OpenDialog contends that its no-code platform empowers insurance behemoths and entities like the UK’s NHS to seamlessly integrate various AI models within a unified automation framework.
According to the company, one insurance client achieved automation of 9 out of 10 tasks, ranging from policy sales to claims processing. Terry Walby, the Co-founder and CEO, emphasized that the platform’s dynamic architecture enables conversations to flow naturally, akin to human interactions, adapting to unexpected inputs without disrupting the dialogue.
Walby envisions that by eschewing reliance on a singular foundational model and embracing a more adaptable platform, OpenDialog can blend different LLMs to offer clients the desired flexibility. While this approach may present advantages, it also raises the question of whether OpenDialog might be perceived as a mere intermediary player rather than possessing a robust proprietary platform.
In response to such concerns, Walby highlights that leveraging multiple language models allows the platform to optimize their utilization for specific segments of a conversation where they are most suitable.
Ed Lascelles, a Partner at AlbionVC, acknowledged the significant evolution in conversational AI since the era of early chatbots and commended OpenDialog for their track record of building and scaling enterprise ventures.
Notably, this venture marks the second entrepreneurial endeavor for the founding team, following the successful scaling and sale of Thoughtonomy, a software automation solution, to Blue Prism in 2019.
