Artificial Intelligence: Slipping Into Everyday Life and Subscription Businesses

By incremental steps, we are tiptoeing into a future in which computers drive our cars, make our restaurant reservations, and, yes, help businesses with recurring payment issues.

Source: Bigstock

Recent advances in artificial intelligence have been amazing. My daughter loved chatting with Siri on her iPod Touch when she first got it; now she takes this remarkable technology for granted. No longer just a scary monster for sci-fi flicks, AI is big business now and getting bigger – $7.3 billion worldwide this year, and soon much more:

(Source: Tractica via Statista)

That’s direct revenue. Respected forecaster Gartner takes a look at the value of the business deriving from AI:

(Source: Gartner, via Statista)

With any fast-growth technology, there will be expected benefits and unintended consequences. But for AI, people have been thinking about the consequences for a long time … and the actual benefits may be surprising. Let’s look at both.

DANGERS OF AI

From Frankenstein to Metropolis to the reboot of Westword, we humans have delighted in scaring ourselves about the dangers of granting intelligence to our own creations. That may be science fiction, but down-to-earth businessfolk are seriously looking into automating and replacing workers in a variety of white- and blue-collar jobs.

Dice Insights reports on AI taking over programmer jobs.

Radiology Business suggests that AIs reading mammograms will improve radiologist efficiency … so I guess we’ll need fewer of them?

Google recently announced AI systems for performing common tasks over the phone, such as placing a restaurant reservation. I suspect that means we will be needing fewer telemarketers in the near future.

Even Forbes is asking, “Do You Fear Artificial Intelligence Will Take Your Job?”

The answer is yes and no – some jobs will be created, and others will go away.

(Source: Bullhorn, via Statista)

But all of that is just the “usual” creative disruption that technology brings. What about the more scary, end-of-the-world stuff? Is that something to fear?

The answer to that lies in the two kinds of AI, one that exists, now, and one that does not, yet. “Narrow” or “weak” AI is the kind we have at present, with machine learning and brute force computing combining to translate languages, read license plates, and predict fraudulent credit card transactions. On the other hand, “general” or “strong” AI  – if it ever exists – will feature a true intelligence that can think, and even rewrite its own code so that it gets smarter and smarter.

For a solid primer on the entire issue, look into this two-part series on strong AI at WaitButWhy. The question is, if we throw enough connections and processors and computing power into a machine learning or weak AI system — say, more connections than are found in a human brain — will a strong AI just emerge? And will it present a genuine danger?

Look again at the remarkable growth forecasts at the top of this column. And consider articles like this one from MIT:

  • Wait-and-See Could Be a Costly AI Strategy — Early adopters of AI will share a global profit pool valued at $1 trillion. Will your company be among them?

In this intensely competitive environment, will AI researchers take every possible care to prevent creating a strong AI by accident?

The fact that strong AI does not yet exist may be reassuring, but a large number of experts do not agree. They think that by the time a strong AI is created, it may be able to improve itself so fast that it will be super-intelligent before we can stop it. What experts? Vox put together a round-up of 17, and there was no consensus on the problem.

AI FOR RECURRING PAYMENTS

But let us step aside from apocalyptic futures and focus on the present moment. What can artificial intelligence possibly do right now to make a difference to a subscription business? Well, in an industry that depends on credit cards for recurring revenue, artificial intelligence is being used now for two important functions.

First, AI is being used for batch transaction processing, which is of particular interest to subscription firms. When a recurring customer’s card is run automatically, often monthly, usually in a batch with a bunch of other card numbers at the same time, the customer is not on the line if the transaction is denied. Perhaps no human is involved at all. So what happens when you cannot ask the client for a different card or other form of payment?

Well, now there are computer services that help with “transaction recycling.” One payment processing company, Elite, explains it this way:

  • Subscription and recurring payment models are being used more often. For different reasons, authorization attempts are not always successful. Making an additional attempt to authorize a card payment can result in more profits. Making an additional attempt automatically can save money and will allow your revenue stream to continue.

You can try non-intelligent recycling, by using a strategy such as retrying at different days of the month. Or you can use artificially intelligent recycling. A data management company, ROI, explains how it works:

  • Intelligent recycling uses special algorithms, developed through statistical analysis across all of their commercial merchant portfolio, to automatically retry declined credit card payments and optimize that process to maximize revenue recapture while minimizing the costs of reattempts. These algorithms understand – by card type – the exact times, days of the week and dates within a month to retry a rejected payment. By using these sophisticated credit card re-try algorithms, a non-profit organization can recover significantly more of its failed payments in comparison to standard reattempt techniques while controlling costs on multiple reattempts.

How effective is this? A presentation from Oxfam, Vantiv, and ROI reveals that order approval rates can reach just about 90% using this AI methodology:

(Source: Oxfam, Vantiv, and ROI)

The implication for subscription companies is that computing power here translates into fewer subscriptions lost to card catastrophes.

Another important present-day application for artificial intelligence and subscription companies is in fraud prevention. Here, the field is more mature, and there are a wide array of applications at work. Here are a few.

PYMNTS reports on an AI system designed to counter microtransactions that criminals use to test stolen card validity.

IDG’s CSO reports on a company named Galileo that uses AI:

  • Galileo’s “standard offering” was a rules-based anti-fraud engine, which it fine-tuned to be more accurate over time. That morphed into a true AI engine that allowed Galileo to be even more accurate and more responsive. The AI isn’t a rules-based engine. Instead it learns on its own over billions of credit card transactions and events. [CEO] Wilkes was hesitant to share the exact details of how Galileo’s AI got so accurate, beyond that it looks at 500 different transaction attributes (such as overall account behavior, physical distance between two transactions, and merchant credibility), and uses previous outcomes to improve precision.

TechEmergence reports on current uses of neural networks that predict fraudulent transactions before they occur.

MasterCard bought an AI outfit a year ago, and later last year, IBM was boasting about how it is helping MasterCard reduce fraud with AI.

And in more remarkable merger and acquisition news, CB Insights reports that nine big companies are snapping up almost all the other AI startups. Check out this graphic displaying the corporate feeding frenzy:

(Source: CB Insights)

For good or for ill, artificial intelligence is developing rapidly not just in credit card processing but in a wide range of everyday applications. This TechEmergence article has many examples. But is the rapid rate of development, and the competition between giant companies with deep pockets, likely to lead to lead to a strong AI disaster?

I did a little original reporting and interviewed an expert.

“Siri,” I asked, “what’s your opinion about artificial intelligence.”

“It’s nice of you to ask,” she said. “Now, can I help you with something?” Hmmm. Sounds like someone is dodging the question.

Insider Take

Organic intelligences are now busy making artificial intelligences – at least, weak ones that can’t think. That does not make them useless; indeed, we’re finding more and more ways to use AI all the time. For now, subscription companies can use AI to improve credit card processing. Tomorrow, we’ll find new uses, guaranteed.

Up Next

Register Now For Email Subscription News Updates!

Search this site

You May Be Interested in:

Log In

Join Subscription Insider!

Get unlimited access to info, strategy, how-to content, trends, training webinars, and 10 years of archives on growing a profitable subscription business. We cover the unique aspects of running a subscription business including compliance, payments, marketing, retention, market strategy and even choosing the right tech.

Already a Subscription Insider member? 

Access these premium-exclusive features

Monthly
(Normally $57)

Perfect To Try A Membership!
$ 35
  •  

Annually
(Normally $395)

$16.25 Per Month, Paid Annually
$ 195
  •  
POPULAR

Team
(10 Members)

Normally Five Members
$ 997
  •  

Interested in a team license? For up to 5 team members, order here.
Need more seats? Please contact us here.