In 2011, Jordi Torras, a former VP of Oracle, realized: companies were not optimizing their search technologies as well as they could. He founded Inbenta to solve this by providing professional consulting services while developing an in-house AI toolkit for search optimization.
Over the years Inbenta moved from consulting to providing conversational AI as a service, including AI-powered chatbots, knowledge management, and search engine tools. This proved to be a wise move — Inbenta’s customer base has grown to more than 250 brands across industries including financial services, travel, e-commerce, insurance, automotive and telecom.
In a sign of investor approval, Inbenta today closed a $40 million funding round led by Tritium Partners, bringing the startup’s total amount raised to more than $60 million. Torras says the new money will be spent on people, process and platform R&D to – in his words – “position Inbenta for the expected explosive growth in the conversational AI space.”
“Inbenta develops a comprehensive platform that tailors AI-driven solutions to different industries and use cases for the needs of all enterprises,” continued Torras. “We have spent more than a decade refining our proprietary and proprietary AI toolkit globally, 35 languages, and have continuously improved them through billions of customer interactions.”
Inbenta offers four main products: Chatbot and Messenger, Knowledge and Search.
Chatbot and Messenger – both chatbots – can be built into existing websites and apps (e.g. WhatsApp, Facebook Messenger and Slack) to answer customer questions. They preserve conversation histories, automatically escalate complex inquiries to human agents or a ticketing system, and leverage automation to handle tasks such as securing bookings, scheduling meetings, and changing orders.
Image Credits: Inbenta
Knowledge similarly provides answers to common questions. But it’s not a chatbot, rather it’s a kind of proactive knowledge base that can suggest content for forms, autocomplete requests and predictively search for information. As for Search, it indexes data from various sources, such as product catalogs, and uses algorithms to deal with queries, auto-correcting typos, and misspellings.
Like Chatbot and Messenger, Knowledge and Search can be embedded in most apps and web pages.
“From a business perspective, Inbenta helps companies automate customer interactions, reduce the need and cost of human intervention, and create an always-on channel for sales, marketing and HR,” said Torras.
You might wonder how Inbenta’s AI was developed in that light some chatbot vendors have a history of training their algorithms on user data without the knowledge or consent of those users. Torras says Inbenta’s AI — which is “curated by a team of experienced analysts and computational linguists,” he claims — is “completely anonymized,” with controls that allow users to delete their data from the platform if they choose.
With its portfolio, Inbenta competes for market share on a number of fronts. In the chatbot space, the company has rivals in Quiq and Ada, which provide customizable, AI-powered chatbot services to brands. Inbenta’s Knowledge product competes with offerings from startups like Sana, while Search takes on Hebbia’s document indexing tool.
Torras declined to disclose Inbenta’s earnings when asked, making it difficult to gauge the company’s traction. Matt Bowman, the managing partner at Tritium Partners, didn’t voice his concerns — not that he would, given that he’s an investor. But in an emailed statement, Bowman highlighted what he sees as Inbenta’s differentiator: “multilingual proficiency that requires no data training and that continuously improves with every interaction.”
“Because Inbenta is highly configurable across use cases and industries, Inbenta enables customers to get an immediate return on investment while having a platform that can meet future needs as usage expands,” added Bowman please.
Dallas, Texas-based Inbenta currently employs more than 160 people. It plans to grow that number to more than 200 by 2024.
