What’s in store for machine learning (ML) and artificial intelligence (AI) in 2022? Is it finally the year for the rise of the machines?
Until not too long ago, AI was just an overused marketing term. Many software vendors who sold solutions based on algorithms and fancy regular expressions branded their stuff as artificial intelligence, even though it wasn’t.
Times have changed, and the market is—in a helicopter view—divided into two camps: vendors who use a predefined AI framework and vendors who create their own. I’m not looking into the pros and cons of each, but what does this mean for the users?
AI Adoption Is Slowed Down by Two Obstacles
First, there’s a financial obstacle. Solutions using real AI can still be called “advanced,” and they come with a certain price tag. But it’s the same old story for businesses when it comes to implementing new tech: it’s all about reducing cost or increasing efficiency. We saw the same with the invention of looms, the production belt, and the increased use of automation in IT. Now it’s just the next chapter, but the calculation remains the same: can we reduce the operating cost of human labour with the new technology? Where’s the break-even point?
IT security is a simple example. How many hours does it take to investigate suspicious behaviour? How many analysts are digging through log files, and how much does it cost overall? It’s a task you can easily outsource to a machine, which will show results within minutes—if not seconds—while humans would need hours.
This doesn’t mean the analysts aren’t needed anymore—they can be assigned to tasks requiring creativity, something AI struggles with.
But as mentioned already, times have changed, and AI adoption has slowly increased. One of the side effects of this is lower pricing and more affordable solutions. It’s no longer just the global players who can afford the latest tech.
In the next 12 months, we’ll see an increased adoption rate, even in the small and medium-sized business (SMB) area. It’s no longer “we don’t need this stuff”—instead, it’s “this could be interesting. Let’s try it.”
Controlling the Beast
This brings us to the second obstacle to AI adoption: complexity.
An AI-based solution, whether it’s off the shelf or bespoke, requires customisation to be successful. This isn’t an easy task, and it usually requires development resources. It doesn’t matter if they’re provided by the vendor as a package deal or taken from in-house talent. Actually, it does matter, as the latter would increase the cost again—see above.
But even here, we see an advanced use of technology, or a mix of different ones, to be precise. Some of these solutions come with what’s called low/no-code interfaces, and everyone who can create meaningful charts can deal with such a system.
What’s left are trust issues. An artificial intelligence solution will make its own decisions.
Like humans, this is a process based on experience, knowledge, and training.
But who provides this training? There’s a reason the EU proposed a regulation of the use of AI in recruitment in 2021.
The bottom line is the more this learning process happens in-house, the higher the trust level.
But it doesn’t end here.
Once implemented, an AI-based solution should do the hard work and improve efficiency while using it.
Most of us deal with some kind of business software during our day-to-day lives.
It could be customer relation management, resource management, warehousing, or maybe lead handling in sales.
All of these come with reports out of the box, and though they may have varying levels of quality, they usually do the job.
But occasionally, we require something particular, and this is where the problems start. Some of these business solutions are dinosaurs and don’t make it easy to create custom reports. It’s not unheard of for a DBA to get dragged out of their usual tasks just to create a custom query and retrieve some data points the CEO requires for a meeting.
Wouldn’t it be much easier if there was a box for free text? We could just write “give me the sales numbers for the Middle East during the last quarter, divided by country,” and it would result in a nice chart within 20 seconds.
This isn’t science fiction—in fact, many of us have used a chatbot already, and that’s what it is.
We just need to improve the quality of the responses instead of “Sorry, I didn’t understand the command.” We’ll get there eventually.
AI- and ML-based solutions will get more affordable and easier to use, which will increase the adoption rate. And 2022 might be a good year for it—no crystal ball needed, actually.
We’re still somewhere in a pandemic-induced time loop, and we need to reduce human interaction.
Robots, you’re our only hope.
About the Author
Sascha Giese is a Head Geek™ at SolarWinds, based in the company’s Europe, Middle East, and Africa (EMEA) headquarters in Cork, Ireland. Giese has more than 10 years of technical IT experience, four of which have been as a senior pre-sales engineer at SolarWinds. As a senior pre-sales engineer, he was responsible for product training SolarWinds channel partners and customers, regularly participated in the annual SolarWinds Partner Summit EMEA, and contributed in the company’s professional certification program, SolarWinds Certified Professional.
Featured image: ©Vegefox