Skip to main content

While some surveys show that people would rather talk to a human than a chatbot, whether they’re shopping online or dealing with a customer service issue, it hasn’t deterred companies from adopting them. A 2019 Salesforce report found that 53% of service organizations planned to use chatbots within 18 months. According to According to Statista, the size of the global chatbot market could top $1.25 billion by 2025, a big increase from $190 million in 2016.

A customer’s satisfaction – or lack thereof – with a chatbot ultimately depends on the scenario and the capabilities of the chatbot in question. Obviously, a chatbot that doesn’t answer basic questions will lead to frustration. Countless vendors claim to have thought through the issue carefully, including Loris. But Loris, which today announced it has raised $12 million in a Series A round, is unlike many in that its software is designed to coach customer service reps rather than respond to customer inquiries. .

“Loris was founded… with the idea of ​​increasing empathetic communications around the world. At the heart of our software’s origin story is [an] An AI-based solution providing de-escalation techniques and language suggesting features that guide customer service agents through the toughest conversations,” CEO Etie Hertz told TechCrunch via email. integrates customer sentiment in real time, every day. We see the future of business winners and losers tied to customer service experiences and their direct impact on the profit chain.”


Loris was spear in 2018 by Nancy Lublin, former CEO of nonprofit social advocacy group Do Something and founder of Crisis Text Line, a suicide prevention organization. Lublin had the idea to take lessons from the Crisis Text Line, which trains counselors to calm people down and guide them through their problems, and integrate them into a system that could help employees and businesses navigate the customer conversations.

The approach proved controversial. In February, after Politics revealed that the Crisis Text Line – a shareholder of Loris – funneled data from conversations with Loris, the Crisis Text Line said it would stop sharing the data and demand that Loris delete all data it had received. (The Crisis Text Line argues that the data was processed “securely, anonymized and cleaned of personally identifiable information.”)

“We derive our information from anonymized, aggregated data that has been removed from personally identifiable information,” Hertz pointed out when asked about Loris’s privacy and data retention policies. “We only retain non-personally identifiable data, so that we can continue to use it to improve our services…We are audited annually by companies such as Deloitte to ensure compliance.”

On this last point, it should be noted that the “big four” accounting firms, including Deloitte (with EY, KPMG and PwC), have been found produce a large number of botched corporate audits. But taking Hertz at his word, Loris trains its systems ethically to guide human customer service agents with suggested language and techniques, leveraging AI to analyze real-time conversation data and provide insights like as the main reasons for conversations that end in low satisfaction.

Picture credits: Loris

Recently, Loris – which is designed to build on existing customer service systems – began piloting the ability to use sentiment analysis to predict when a customer might unsubscribe and recommend the appropriate strategy. “We collect feedback from our users and empower agents to help improve the system,” Hertz said.

Look forward

Loris is positioning its platform as beneficial for agents — not just customers — at a time when the volume of customer service requests is increasing and customers have higher expectations of brands. According to a recent Zendesk report25% of customers said they have used live chat more during the pandemic, while 62% say they prefer resolving issues over the phone.

Some to research shows that customer service agents are wary of AI and automation tools. But Hertz argues that, especially for agents whose first language is not English, Loris’ sentiment analysis tools can help them adjust their responses to hard-to-discern tones, leading to better results – assuming these tools are not biased.

“We analyzed tens of millions of customer service messages across multiple domains. This gives us a large and diverse data set to help mitigate bias,” Hertz said.

Brands considering adopting Loris will need to weigh the pros and cons of competing solutions like Google Cloud’s Agent Assist and Contact Lens for Amazon Connect, which also use sentiment analysis. Call center service providers such as Dialpad also offer sentiment analysis features, as do companies including Cogito, saygent and SugarCRM.

Hertz says the plan is to invest late-cycle capital in research and technology as part of a “robust pipeline of expanded functionality” for the “thousands” of agents across a customer base that includes Freshly, Fiverr and Slice.

“These [features] will further enable non-technical managers who oversee customer service and support teams to efficiently and effectively scale the “human touch” of their departments… Specifically, we will continue to evolve our analytics tools — [which] serve as an out-of-the-box data science partner for CX managers to identify areas for improvement – ​​and further augment their current empathetic language guidance with configurable parameters,” Hertz added. While there are other AI companies working on automated responses, they tend to steer towards self-service… [we] focus on meeting the analysis of customers’ intentions and sentiments during and after the conversation.”

Loris, which has 15 employees, aims to triple its workforce this year. Bow Capital and ServiceNow co-led the Series A alongside existing investors Floodgate and Vertex Ventures.