Augmented intelligence versus data literacy - conflicting goals? Qlik's CTO says no.

Profile picture for user jreed By Jon Reed November 2, 2020
Qlik's user conference landed at an important time for the future of analytics. On the one hand, we have Qlik's longstanding commitment to fostering enterprise data literacy. On the other, the pursuit of AI-powered "augmented intelligence." At stake? Only the resilience of our economy.

Confused business team holding a question mark sign

I have a pretty grouchy view of enterprise software acquisitions, including private equity moves, like Thoma Bravo acquiring Qlik in 2016. Too often, what follows seems like a move to squeeze return from an asset, rather than deliver a better product for customers. I'm not sure we can say that in this case. Since being acquired, Qlik itself has gone on an acquisitions binge.

The time to bolster the product is right. Analytics keeps shifting, from on-prem to cloud - and now buzzphrases like streaming data and augmented intelligence keep popping up. Even the deepest analytics vendors didn't historically provide users with "next best action" type functionality. Time to bear down on development, or acquire, or both. Qlik did both. Consider:

These moves gave Qlik customers/followers/prospects plenty to digest at Qlik's Data Revolution (virtual this year). The Americas event took place on October 29th - I expect it to be available on demand soon. In the meantime, you can register and check out other regional Data Revolution events on demand, a nifty feature.

Augmented intelligence versus data literacy - conflicting goals?

Prior to the event, I spoke with Qlik CTO Mike Potter. My burning question for Potter: how does he reconcile Qlik's long-time emphasis on data literacy with machine intelligence, and, now, a so-called Cognitive Engine. Don't we run losing our data literacy edge as we invest in so-called "intelligent" tools - tools that do the work for us? To respond, Potter took a look back. As he told me:

For many previous generations of business intelligence and analytics, it was very much the privilege of the data-educated. They got access to the data, got the opportunity to build reports, dashboards and the like. Through the democratization of data, what we're seeing now is that more and more people need data to do their jobs - without necessarily being educated in data. And so the challenge has been: how do you bridge that gap?

Potter believes tools like Qlik's Augmented Intelligence can actually bring literacy to the forefront:

One of the greatest failings of a lot of the products out there is: there's too much of an assumption of your level of data education. What we've been trying to do with this idea of data literacy is: there are people who know what their data is, but don't necessarily know what a join is, or an ETL process. So what we've been trying to do through our augmented capabilities, is bridge that gap for them.

Tweaking and managing data slows analysis down:

The most effective use of our cognitive strategy and our augmented capabilities is to remove the complexity, the science of data, and just focus on the value of data for the customers and the user.

Some analysts might be excited about shiny new AI offerings; what I like to see is more vendors taking responsibility for that complexity Potter is talking about - as well as the non-trivial mundanities of data cleansing/quality. Potter responded:

For many years, the data life cycle has been a very IT-centric concept. What I believe is happening, through the democratization of data, is that it's now being more distributed and enabled throughout the entire organization.

But if you want to get there, you must overcome the IT/business data gap:

I often like to characterize it by a simple Venn diagram. A lot of IT organizations haven't managed data. However, the business has a need for data. If you think about the intersecting set between those two Venn diagrams, you can often challenge many a CIO or Chief Data Officer as to what that intersecting set is.

Often, they don't have the visibility in terms of how relevant that data is to the business at any point in time. By contrast, the business's needs for data, and the problems that they're trying to solve with data often missed - or they can't be supplied by the data that we currently have and manage. And so from my perspective, the opportunity is to create a closed-loop life cycle around both perspectives, so that the consumer of the data can actually drive requirements into data strategy, as much as the sources of data.

Let's say an enterprise buys into this approach - doesn't mean it's easy to bridge those data gaps. What is Potter's advice for companies seeking to propagate data literacy?

I think the simplest place to start really is being able to do a gap analysis in the context of that Venn diagram that I described. Many organizations spend a lot of money housing and managing data without knowing when it's being used - or if it's relevant... You can go through a fairly definitive process that can actually help you do that.gapping. 

Then, Potter says, you can manage those two data extremes - governance on the one side, and "analytic freedom" on the other:

You do that in the context of what the latency requirements are, what the relevancy requirements are, and the lifespan. If you get all that right, then you have a system that can provide the governance - while also providing analytic freedom for the community.

My take

When enterprises assess economic scenarios, there are plenty of difficult unknowns. As they look at issues like return-to-work, or whether to re-open storefronts, they're making assumptions around in-store demand, perhaps - or what constitutes a safe environment, based on what we know at this time.

Some of those assumptions may be correct; we have learned important things about infection spreads, testing, and contact tracing, but there is much we don't fully understand. No, analytics won't solve this problem. But increasing data literacy should help companies keep their workforce engaged with the right problems.

Qlik thinks they have another useful angle here, via the conversational interface of their Insight Advisor Chat. Does giving users a voice-based way to query add significant value, or is it just bells and whistles? Potter says that so far, it's been very well received by customers. Certainly, enterprise search is a time-waster, but can a conversational UI go further than (potentially) saving time? Can it lead users to fresh insights? Potter says that answer is an unequivocal yes. I'd like to talk to Qlik customers directly about this, but I wouldn't accuse Qlik of analytics-as-usual. They are pressing the issue, now let's see where they get further traction.