When Facebook bought Ozlo, a start-up that applies “artificial intelligence” toward better understanding of text-based input, it signaled a deep understanding of the challenges posed by conversational commerce. Ozlo emerged last year as a cross between Siri, Yelp, and Facebook M. Users could inform Ozlo of dietary restrictions, for example, and have it recommend top rated restaurants that could accommodate suitable restaurants with top ratings. Ozlo leveraged deep integrations to partner sites, in order to save its users from having to open multiple apps to locate the same information.

According to Associate Editor Tess Townsend in Recode, Ozlo had received $14 million in venture capital funding as of last March. The company’s consumer-facing natural language mobile assistant launched just over a year ago. This spring the company apparently pivoted when it began offering its natural language understanding and data analysis tools to other companies building bots of their own.

While there’s plenty of competition in the bot tools space, Ozlo claimed key differentiators that support better conversations between humans and virtual assistants. Chief among them, its technology has the ability to recognize when clarifying questions need to be asked in order to provide accurate answers or complete results. Most virtual assistants stumble when they don’t have a tailor-made response to a specific question. Ozlo apparently relies on internal knowledge ontologies to understand the general category of a question so that the assistant can prompt the user for further clarification until it finds a suitable response.

Charles Jolley, CEO and co-founder of Ozlo, also noted that Ozlo’s knowledge model makes room for “probabilistic assertions alongside accepted facts.” For example, most virtual assistants can answer a question about when a restaurant opens. But the vast majority don’t have a response when you ask if the restaurant is good for kids. Ozlo is built to understand and field both factual and subjective inquiries.

The Ozlo tools and team members will become part of Facebook Messenger. While it’s not entirely clear how Facebook will use the Ozlo technology, the acquired capabilities could enhance Messenger’s text-based bot framework as well strengthen Facebook’s M assistant. Reports from earlier this year suggest there’s been disillusionment at Facebook with the viability of truly conversational interfaces. The acquisition of Ozlo signals that Facebook is ready to bake better understanding into its “bot” ecosystem to make text-based virtual chat more appealing and effective.

Categories: Conversational Intelligence, Intelligent Assistants