Lang.ai looks to help orgs extract value from customer conversations, with AI

Turning conversations – from customer support requests to user feedback – into tangible business value is no easy task. It’s also an ideal use case for AI-based automation.

Among the vendors helping organizations use AI to derive value from customer conversations is San Francisco-based Lang, which announced today that it has raised $10.5 million in a series A round of funding. Lang’s platform integrates with help desk, customer relationship management and user-facing operations for feedback and requests. The system uses an unsupervised learning model to adapt to the constantly changing flow of information by categorizing data and then helping to determine what should be done with the data to help improve user experience and business outcomes.

“There has been a growth in the volume of conversations that business teams have to deal with, especially things like customer support, which has been accentuated during the pandemic,” Jorge Peñalva, CEO of Lang, told VentureBeat. “Sure, there are a lot of AI technologies, but in general, they’ve been built by engineers for engineers – so they have a lot of complexity. We believe there should be a better way for business users to use AI.”

Lang certainly isn’t alone in its corner of the market. Zendesk, for example, has built out its AI capabilities in recent years to help with its customer service platform. A core element of its capabilities came from the company’s 2021 acquisition of Cleverly.ai.

CRM giant Salesforce is also very active in the AI ​​space with its Einstein platform. Contact center technology vendor Genesys actively continues to grow its AI capabilities with its Google partnership.

A recent report from Fortune Business Insights estimated the size of the global customer experience management market at $11.3 billion in 2022. The report forecasts the market to grow at a compound annual growth rate (CAGR) of 16.2% over the next seven years, reaching $35.5 billion by 2029.

How Lang uses AI to derive value from conversations

Peñalva is keenly aware of the market potential and the competition. In his view of it, Lang provides a differentiated approach thanks to the use of an unsupervised AI model.

A common approach to enabling AI is the use of a supervised model that trains against a given set of data. The challenge with the supervised model is that AI is often trained on static data. Peñalva noted that data changes quickly and for organizations to truly be responsive to users, training on static data isn’t good enough. That’s why his company developed a purpose-built unsupervised learning model which is constantly looking at data that is constantly changing.

How it works: Lang connects to the customer data and the unsupervised model analyzes the data, transforming it into simple “concepts” – which Peñalva explained is a business term for an item or operation that a company needs to track. A concept could be a delivery date, a product, or a credit rating, for example. The AI ​​model extracts the key concepts in a conversation automatically, so they can be grouped into categories that make sense for a particular business.

The interface to the categories is provided to users in a no-code model, enabling an organization to group things as required. The no-code interface also helps to provide a form of explainable AI, so users can easily see how the unsupervised model extracted concepts and which categories the concepts are placed into.

Scaling operations

Using AI to derive business value from conversations can also help organizations to scale operations.

One example is with Lang customer Ramp, which provides online tracking services for spending. According to Peñalva, Ramp’s challenge was that it wanted to quickly scale up operationally. With Lang, Ramp was able to more rapidly categorize customer requests into categories and then provide automated workflows to accelerate resolution. For example, Ramp can make sure that an inquiry about a credit issue is routed to an agent that can respond quickly to that type of request.

Ramp also uses Lang to understand customer feedback. As Ramp builds out new products, feedback and requests are analyzed by Lang to better understand how the new product is being received and what if any changes need to be made to optimize user experience.

“We really operationalize their support data for automation and also for internal insights that other teams can use,” he said.

With the new series A funding in hand, Peñalva wants to continue to help organizations more easily derive business value from data and help them to automate repetitive tasks.

“We think a lot of companies are gonna be thinking these days about how they become more efficient,” he said. “There are a lot of inefficiencies when you think about the repetitive tasks that people are doing in their day-to-day jobs, when they really should focus on more high-level tasks,” Peñalva said.

The new funding round was led by Nava Ventures and included the participation of Oceans Ventures, Forum and Flexport Fund.