Ideas about data democracy and “analytics (and BI) for the masses” started gaining currency in 2004 after second generation BI software players Qlik and Tableau arrived on the scene. These companies certainly advanced the cause of data democracy by making BI easier to use and more visual. So how is it that thirteen years later BI and analytics adoption rates continue to flatline at a mere 21 percent? Where’s the democracy when 4 out of 5 users are abandoning these applications?
Like one of its birthplaces, Rome, democracy wasn’t built in a day. Early democracies excluded most citizens from voting. Even when new groups of people won the right to vote, power wasn’t distributed evenly, favouring those with wealth and land. Likewise in business, despite the fact that more people today have access to analytics software, lingering issues with ease of use and trust in data have limited engagement among the masses.
It’s only when a front line business person gets easy and trusted access to analytics, that they begin exercising their ‘vote’. It is with their interactions and engagement that they contribute to the collective wisdom of the company – creating new insights and demanding higher data quality, new data sources, and faster availability. Today we are on the cusp on this exciting reality.
AI will bridge the data access gap
The last hurdle left to clear is the “data access gap” between the supply and demand for data insights. On the demand side of this gap, mainstream awareness of data’s potential is turbocharging the desire for analytics among business users on the front lines in sales, marketing and operations. On the supply side of the gap however, data’s proliferation and complexity is making it increasingly cumbersome for all but the most savvy data scientists to get at it and analyse it.
It’s early days, but AI is already starting to bridge this gap, enabling business people to make easier and faster data-informed decisions. Front line employees are now doing things reserved only for data analysts thanks to new AI-powered features that promote continuous learning and a much easier way to tap into collective knowledge. The innovations that follow are being applied in analytics and have the potential to finally empower the corporate frontlines and transform the workplace.
Content from our partners
Machines that understand humans
Natural Language Processing (NLP), which allows machines to interpret and answer questions in human language, has actually been under development since the 1950s. NLP’s recent application in systems like Alexa and Siri showcase its ability to deal with simple, unambiguous questions like what the weather forecast is for a particular day. (It gets a little trickier when you ask these systems questions about, say, the meaning of life!)
Combining NLP with Relational Search is a lot more powerful and suitable for analytics. This enables someone to pose a question in a Google-like search bar on a data set, even if that question isn’t specific enough to yield a perfectly accurate answer. When this happens, guided suggestions help the user refine the search until it is specific enough.
This means business users no longer have to go through data specialists – sometimes waiting days or weeks between iterations of queries – to get insights they need. The beauty of this AI-driven capability is that both the system (by gaining more context) and the user (by learning about the available data and how to construct good queries) are both getting smarter at the same time.
Recommendations get personal
Personalisation can enhance analytics even further. This works in a similar way to YouTube, which applies sophisticated algorithms to show users recommendations based on the genres they like, what’s popular in those genres and how they’ve rated other videos. In business analytics, users are shown recommendations of data insights that are popular based on what they and others in their teams or their roles are interested in. Again, the machines and the humans are getting smarter over time. ROI from these systems continues to grow exponentially.
The “Report Factory” gets automated
AI-enhanced systems take care of the heavy lifting involved with analysing massive, diverse data sets and drawing conclusions about data trends, groupings and anomalies. Even if millions of skilled data analysts existed in the job market and companies could hire them, humans using old BI systems can’t produce results fast enough to support real-time business-wide decision making. Another place where AI helps plug a chronic skills gap is in data visualisation. Algorithms can automatically select the most appropriate chart for representing a specific data set that also matches the end user’s preferences. Finally, machine-driven data profiling is starting to automate aspects of data preparation and hygiene, which are mind-numbingly boring, time-consuming, and error-prone tasks to perform manually.
Your data will talk to you
Natural language generation (NLG) involves machines translating tables of data into spoken words that humans can understand. Imagine that instead of having to view a dashboard with the same limited slice of data every week, your system recommends insights for you based on existing data characteristics, your role and objectives. Imagine further being able to listen to this report in your car on the way to work. That’s what NLG has the potential to offer.
Data democracies will dominate
It’s no coincidence that the world’s most advanced democracies are also the most prosperous. Democratising analytics in business will also lead to financial gains through higher profit margins and greater efficiency.
But this isn’t just about ruthless business efficiency. AI’s ‘virtuous circle’ in the context of business analytics is one in which people and machines help each other get smarter and focus on the tasks they do best. In this way, AI will help more corporate citizens feel that the work they do is more valuable and fulfilling. I strongly believe this will soon lead to much higher analytics adoption rates than the current 21 percent. And when that happens, we’ll finally know that “true” data democracy has arrived.
