Our Latest Articles and Insights | RESULTS.com

How AI Will Affect Your Business Within 5 Years

Technology has already changed business completely. Everyone knows that.

But the real revolution is yet to come.

Businesses who prepare for it now will have outstanding advantages in five years.

That’s because Artificial Intelligence (AI) is on the verge of exploding into the mainstream.

And there’s only one way for small- and medium-sized businesses to benefit.

Better than H.A.L.

But before we get to that, let’s look at what AI actually is, and what’s happening in the market that makes me think a revolution is right around the corner.

First of all, nothing like H.A.L. in 2001: A Space Odyssey exists today, nor will it come to exist anytime soon. H.A.L. had truly artificial consciousness (and his own agenda.) 

The first version of AI mimics intelligence, which is incredible enough. But it doesn’t actually become intelligent.

It's called "machine learning," and it does something far more useful.

Let me give you two examples that will explain.

First, there’s the computer that beat the world Go champion. Go was invented 2,500 years ago. It’s played on a 19x19 grid, and the pieces are black and white stones that you place on the intersections in an effort to capture territory.

The rules aren’t important. What matters is that the computer that beat the world champion was made of eight “GPUs,” which are off-the-shelf processors originally designed to display graphics.

It learned how to win the game without human help. 

All the humans did was hand-code the rules into AI software. Then the computer played against itselfmillions and millions of times.

That’s where the learning comes in. The computer doesn’t “understand” what it’s doing. It just develops a set of probabilities: “If X happens, and I do Y, then I have a better chance of winning.”

That’s AI as it stands today.

Here’s another example.

A company not far from our San Francisco office, MetaMind in Palo Alto, used a similar technique to get a computer to spot prostate cancer by looking at all sorts of test results. 

Again, the computer has no understanding of what it’s doing. But it was able to find patterns that enabled it to analyze the data and make predictions better than humans could.

The Force Awakens

Salesforce.com just bought that company for an undisclosed amount.

Another AI startup that focused on voice recognition was recently acquired by Facebook. Google bought one named DeepMind Technologies.

When off-the-shelf computers can learn Go better than the best human, and when companies like these race into the space, you know that the field is rapidly evolving.

That’s why it’s so critical for businesses to pay attention now.

Because all of these AI projects have one thing in common. 

The main thing that AI needs, more than anything else, is a massive set of data.

Since all the machine can do is find probabilities, small amounts of data won’t work. Say you flip a coin twice and it comes up heads and then tails. If that’s the whole data set, and you feed it into an AI program, it will predict that the next flip will come up heads. Heads, tails, heads, forever. It can’t “understand” that the flip is always random. All it can do is look at a tiny pattern and come up with the wrong answer.

On the other hand, if it plays millions of games, or looks at millions of test results, it’s almost at the point where it can actually make better predictions than people.

The Human Element

Including human input in this process will make it work even better.

Google was doing this with what history will probably consider to be the first “proto-AI” to actually work.

They combined what humans could do with what computers could do. And it was all about collecting data. Google would show you a list search results by matching the “keywords” in your search to what was on a web page.

And every time someone clicked on a result, that person added value to the company.

By clicking, Google learned from that person which one of the search results was the best one. It stored this human knowledge in a vast, ever-evolving database that included search patterns and all the properties of all those web pages. It studied that database continuously. Now the company’s at the point where it can often correctly guess what you’re looking for before you’re even done typing.

That’s the sort of partnership we will soon see between AI and business leaders.

AI will look at your data and it will make predictions. It will make recommendations. It will issue warnings, it will suggest opportunities. You’ll judge them all. You’ll follow some – and ignore others.

Then, time will tell. Your database will grow, and your AI will get stronger.

How to Start Now

Five years from now, the companies who have this will be unstoppable.

So how do you make sure you’re one of them?

First, remember that the world champion Go player only has eight processors. That’s a pretty strong indication that AI will be affordable.

Second, look at the data you have around you now. Salesforce might already contain an incomprehensible heap of it. Same with finance, and with project management, and with marketing, and so on.

Think about the most important intellectual property you have, which, truth be told, is probably a bunch of Excel spreadsheets where you do your most creative work.

Everyone has that sort of "hard data," and it will certainly be needed.

But that's not enough to truly take advantage of AI. 

In fact, the most important data will be the sort that only the best business management software solutions collect today.

It’s not just sales figures and time lines. It's not a bunch of colorful charts.

It’s the human element. 

Besides hard data, this is what you need to collect today if you want to use AI to compete tomorrow:

  • What processes drove each specific set of hard data? 
  • How was the goal defined? 
  • What projects were required? 
  • How did they fit together? 
  • What tasks were involved? 
  • How did you track progress? 
  • How did you define success?
  • How did you succeed?
  • Which people contributed the most? 
  • Over what time frame?

In other words, you need to capture data about the people who did the work. You need to know how your goals were defined, and how your people pursued them. You need to capture data about the questions they faced, and the decisions they made.

The businesses who have captured this data will be able to use AI far more effectively than those who only have "hard data."

The human element is unique to you. It’s unique to your business. It's unique to your team. If you've collected it, then when AI becomes mainstream, you will have a far richer – and far more meaningful – dataset than your competitor.

You will start the AI game with a huge lead. 

As Scientific American put it, “Machine learning does not create information; it gets the information from the data. Without enough [data], machine learning will not work.”

To succeed tomorrow, you need to start collecting this data today.