Stockholm startup Stravito is today launching a new offering to enterprise search. Using proprietary generative AI allows users to ask full questions using everyday language and receive short natural language answers synthesised from various proprietary sources.
Stravito is an award-winning Swedish SaaS Scale-up whose AI-powered enterprise insights platform allows the world's largest companies to easily store, discover, and integrate market and consumer insights.
Using the Stravito platform, global organisations such as McDonald's, Comcast, Burberry, and Danone put more insights into the hands of employees worldwide, so they can maximise the value of their market research and insights and make more informed business decisions.
I spoke to Stravito founder and CEO Thor Olof Philogéne to learn more.
At the core of brand messaging is using datasets to tailor communication to internal and external audiences.
Philogéne explains, "Market research helps brands understand and account for what makes your customers tick. Then you apply that understanding to adapt your communication to match the customer's inner wiring and worldview."
Generative AI can prompt brands to create different versions of brand messaging sensitive to specific audience segments. And it begins with asking questions.
It's easier to ask the right questions
Using Stravito's new enterprise search, upon asking and conversing, answers are returned with additional "related questions" that offer follow-up search inputs based on the current conversation.
According to Philogéne
"Generative AI has shown us that finding the best answers requires asking the right questions, so we're making it easier to find the right question.
Users can further refine their search by including related questions alongside each answer to find the most useful answers, providing critical advantages for urgent needs and long-term business success."
Users can also dive deeper with recommended questions – all with the security of a closed system that generates accurate answers exclusively from a client's owned internally owned data and research materials – unlike open AI apps that pull from the public domain.
Reducing Generative AI hallucinations
Stravito's exclusive use of respected and vetted data sources removes the possibility of random information noise clogging up the research process or even worse, delivering false insights.
Philogéne cites that challenge of what he calls "prompt dependency and trust":
"If you get it wrong in the prompt, you're still very likely to get a highly convincing answer. How it can blend correct and incorrect information makes it even trickier. Human values are human emotions, often requiring a nuanced perspective.
So not all questions have a single right answer. And when you try to synthesise large sets of research reports, there's a risk that some key details will fall through the cracks."
In response, every search using the platform also returns clickable links to source documents, so users can feel confident in the validity of the provided answers. This way, the algorithm provides context and identifies potentially contradicting data points, offering a nuanced and thorough approach.
Philogéne asserts: "The updated platform ensures maximum accuracy by exclusively utilising and searching the client's own stored data – nothing more, nothing beyond."
Private AI helps increase security
Stravito's technology runs on approved, secure enterprise cloud environments which are already in use for applications like SharePoint Online.
Of course, Generative AI is not without risks. and Philogéne contends that "leveraging an AI engine provides incredible power and speed to parse large volumes of data, and that comes with great responsibility":
"Some public releases of AI software have had unexpected consequences, highlighting corporations need access to private services and functions that protect their IP and market research.
We've put in the time and work required to deliver an enhanced user experience while assuring the safety and privacy of all data and activity in the platform."
Philogéne also asserts that startups interested in leveraging AI should approach it from a desire to solve persistent end-user problems. He suggests identifying a couple of really hairy problems that either customers can't solve, or are just very demanding and seeing how you can apply it."
Stravito's new generative AI functionality will be rolled out in stages to enable users better to leverage their company's available data and research.
Stravito is already developing additional generative AI-based workflow enhancements to further revolutionise its insights platform beyond this initial offering.