How genuine is Artificial intelligence?

Vast Data went from $1 million in annual recurring revenue (ARR) to $100 million in three years of selling because, it says, it’s designed for the age of artificial intelligence (AI). Hang on though, aren’t margins allied to mystery? Wasn’t AI designed to demystify everything? Well of course not. That’s just what insiders say to keep the layman bamboozled it appears.

Channel experts have long subscribed the IT industry is a fashion business. There’s high profit to be made from AI while it’s in its demi-monde phase of cloud couture. With demand outgrowing supply there’s plenty of time for the bland to Lead the blind and make a fortune and invest it before the next banking collapse. However, some ancient principles, such as garbage in equal garbage out, will always apply. Artificial intelligence is only as good as its genuine intelligence, its data, and if you feed a machine on outdated shibboleths it’ll come to some very odd conclusions.

Global enterprises, says Renen Hallak, founder and CEO of Vast Data, are looking to deploy innovative, yet proven technologies into their data stack. One of the worrying aspects of the IT industry is that everyone seems to have taken standard English language terms, such as ‘agnostic’ and applied their own proprietary meaning. So they’re training machines on pure rubbish. It must be confusing for a machine to read the phrase ‘platform agnostic’. If the machine consults the Oxford English Dictionary it will find only one definition of the word agnostic, and therefore conclude that it refers to someone who doesn’t believe in the existence of computers. (Any person that doesn’t believe in computers, the learning machine might reason, is in the wrong job, surely). The same confusion must arise when any AI/ML comes across a description for an ‘explosion’ of data.

Still Vast data must be doing something right, because it has built the data foundation for Nvidia’s AI superclusters. Then there’s the work it does for Veeam, Vertica, Commvault and Dremio. It has quadrupled the number of Fortune 1000 customers, with its top one hundred new customers spending more than $1.2m on average. Now it’s looking for partners in the UK.

Luckily, there are others in the industry that can also talk about the future in terms we can all understand.  Brian Dunleavy, commercial director at channel partner Viadex, knows enough about the subject to explain it in simple terms, which is a rare talent these days.

“I run a tech business and secretly many of our consultants have admitted to the fact that complex language is a legacy behaviour that keeps technical people occupied,” says Dunleavy,

Data comes in all shapes and sizes and lives in different homes. The more relevant data you can access instantly, the more accurate and instant your decision making becomes, Dunleavy says. Solving the riddles of your industry data is what makes you more useful than your rivals. Using AI, they are able to make instant but comprehensive business decisions that have a lasting impacts on their future direction and strategy.

At the end of the day, don’t most businesses just buy a computer because it saves them time and money? Saving time and money are the minimum expectations of any technology, says Dunleavy. You should be doing better on live inventory decisions, investments, client pricing and offers, revenue and profits. The challenge is you have historic data in legacy systems and silos, you have current data and you have predictive data and modelling. But they are all meaningless unless you can find, interpret and present relevant conclusions instantly. Otherwise you are like the First World War generals who sent thousands of troops to mown down in a hail of bullets because they they’d be trained for wars where machine guns didn’t exist.

Vast’s conflict winner is that it gathers the latest intelligence from every corner, even existing, legacy hardware investments.

The likelihood of Vast winning the data race is heightened by its relevance, timing and cost. All of those variables have past, current and predictive elements, says Dunleavy. AI and Machine learning algorithms make this happen instantly, across past, current and future predictive data, such as weather, crop yields and economic indicators. VAST algorithms can static, live and predictive data, regardless of the machine it currently sits on. Getting to see it instantly is the trick, as long as it’s relevant and accurate enough to give users what they want, how they want it, when they want it. That’s where the margins come from, Dunleavy says.

Yes, I think I can understand that.