More business intelligence projects are being started by the business rather than IT, and Tibco’s analytics head, Brian Gentile (pictured), believes BI needs to adapt for mass-market adoption.
Gentile is senior vice-president and general manager of Tibco's analytics product group. Computer Weekly met up with him in London during his recent European visit.
He says genuine self-service will open up BI tools to be used by far more people than is currently possible.
Gentile was previously CEO of Jaspersoft and joined Tibco when the company was acquired last year. He now leads the analytics business, which includes the Spotfire analytics tool and Jaspersoft BI suite.
"Self-service techniques have to take a leap forward in order to reach a much larger market," he says. "It is one thing to make a tool easy to use; it is another to draw people into an analytic conversation."
As well as being easier to use, Gentile says BI tools should be integrated into the way people work. "Most people won't even use the tool; they will not want to step out of their work environment to analyse data separately," he says.
"They need analytics to be brought to them inside the business applications they are already using. Analytics becomes a thing that you do, not a place where you go."
Read more on analytics
Merging structured and unstructured
Speaking about the trend to merge unstructured and structured data, Gentile says: "In certain very specific business applications, we are seeing the clever use of unstructured content captured in a variety of forms with quantified, traditional analytics."
Examples include consumer packaged goods and retail, where sentiment information from customer support lines and social media channels are constantly being analysed. Gentile says this content is then quantified using word trees, graphs and charts.
The data is then compared with store-based data and regional information, and Gentile says combining these two data sources creates news insights: "The retailer can analyse what is driving consumer confidence."
Looking further ahead, he says: "I have had conversations with organisations that would like to use intrusive techniques, such as scanning everyone's emails to gauge employee satisfaction across the organisation and potentially glean product ideas."
The challenge from a BI perspective is how to visualise such data to make it meaningful, says Gentile. He expects it will be a number of years before the combination of structured and unstructured data produces killer applications for creating new value.
Intelligent agents will power consumerisation
Greater automation in BI will pull more people into the analytics conversation, driving greater consumerisation, says Gentile. "The more we can do in software to guide someone towards a better analytic experience, helping them to be better analysts, will help them make better decisions."
The more we can do in software to guide someone towards a better analytic experience will help them make better decisions
Tibco provides a recommendation engine within Spotfire that guides users to the most appropriate form of visualisation for the data being analysed. Gentile says: "There is no reason why the same heuristic machine-automated approach could not be developed further. We should be able to recommend the data sources to analyse."
Gentile adds: "The business intelligence sector is heading to a point where an automated personal intermediary could one day work as an agent looking for analytics in line with your needs, matching up questions you are likely to pose with data sets that are available."
Such agents could work using natural language queries, as in the Siri and Cortana voice-based user interfaces on Apple and Windows devices, he says. It would be equally viable for the agents to be set up in the user's personal BI portal.
But combining structured and unstructured data in a meaningful way is pushing the limits of machine intelligence.
"It is all very well to say I already have the quarterly sales broken down by region, but it requires a different level of intelligence to say I will have to attach to another data source to query a new set of tables," he says.
Today it takes human intervention to combine unstructured data results with structured analytics, but Gentile says automated processing is improving. For instance, the most recent release of Spotfire has a content analytics engine, he says.
Gentile says this represents the first step of a journey Tibco has begun where patterns of data from an associative engine can be used to analyse content side by side with structured data and present the findings in a dashboard that is constantly refreshed.