A.I. is Dead. Long Live A.I.

A.I., as journalists understand it, is dead.

And I could not be happier.

The New York Times, the Wall Street Journal have each written its obituary within the last couple months. The NYTimes called IBM Watson a “sobering example of the pitfalls of technological hype and hubris around A.I.” Of course in the same story, they paid the Allen Institute to review the system and found that, its actually very good at understanding human language, and sorting through vast piles of data.

But, no, Gray Lady. It can’t cure cancer.

Then, the Wall Street Journal wrote “there’s nothing ‘intelligent’ about AI at all” while going on to say venture capital funding for AI nearly tripled between 2017 and 2020. And investment in AI is even hotter in 2021, exceeding all of 2020 —  nearly $30 billion by July 15^th.

But, no. It can’t replace humans.

Working with AI for the last few years, I thought I would be upset at the stories, calling this technology failed and listing the road kill of unfulfilled promises along the path that led to where we are now — as measured by the newspapers that announced them. Instead, I found myself relieved.

“Good,” I thought. “Now I don’t have to explain that AI is not going to turn your MSP into Accenture by 2025. You will not be able to fire your entire help desk and replace it with a smooth, clean running UX. You will have to put work into it. Data into it. People will need to hold its hand sometimes and help it find the way.”

The reality, as shown in both stories, is that AI is not a magic box that with the right amount of tinkering can solve the ills that plague mankind. It is as flawed as its creators, and as imperfect as all technology, but that doesn’t mean its dead, or useless.

Far from it.

Right now, the national labor shortage is hitting MSPs particularly hard since they need technical workers to fulfill  obligations to their customers under existing contracts. To honor those deals, business owners have to make one of two unpleasant choices: hire someone who is not qualified, but might be trainable, or hire someone who is qualified, but comes at a cost that is greater than their value, vis a vis the contracts.

In the first case the risks is that the person cannot learn, and in the second that the juice isn’t worth the squeeze. But both of those business risks are mitigated through the same technology that so disappointed the New York Times and Wall Street Journal: AI, and specifically Watson.

A new help desk employee – whether they are an IT veteran, or a rookie — enhanced with IBM’s Watson through CrushBank can work faster than the more experienced person next to them. We see it in the tickets they close, the problems they solve, and the resolutions they can find, qualities their peers who rely on muscling through tickets cannot match.

With AI, businesses can open up their entire service history to technicians, so when a ticket comes in, workers can instantly see how the organization has solved this problem before, and they do not have to reinvent the same solution again.

There is a common fear that AI is going to replace human labor, but no one mourns humanity’s failure to capture its own effort. AI may someday, in some iteration, replace humans. But right now it cannot do that.

What AI is doing right now is rescuing the hours worth of work, and decades worth of solutions that MSPs have created through the sacrifices they have made to run their business. The work that someone put into a customer’s problem years ago, very likely has value for another customer in the future. Yet, now, when that same problem comes in again – and it will – the path of least resistance, for most engineers is to recreate the wheel.

So, the danger for MSPs is not that AI will replace humans. Rather without AI, and access to its past solutions, businesses will allow “computer problems” to consume their labor costs. It is not “machines replacing people,” but machines overwhelming them.

If we aren’t learning from our past, and we are in fact, repeating it, let’s at least get the answers faster this time.