6 Critical Questions to Help Businesses Cut Through the AI Hype

Don't buy an artificial-intelligence 'solution' just because you can. Ask how AI's strengths align with your company goals and supplement other resources.

learn more about Brett Jackson

By Brett Jackson

Tom Merton | Getty Images

Opinions expressed by Entrepreneur contributors are their own.

It's always gratifying when experts confirm what you've suspected. Research firm Gartner put a shot of reality into our morning coffee this past summer with its critical analysis that artificial intelligence (AI) had reached the "peak of inflated expectations." Frankly, I think a lot of technology vendors are blowing smoke about their capabilities.

If your expectations were for Elon Musk-worrying AI, dial them back to the level of business software. It's actually less mundane than it sounds. AI already has changed how we use data to make sense of our world. It's also become a corporate fashion. For every AlphaGo Zero there are a thousand firms -- startups and established companies -- sticking the AI label on their wares like pinstripes on 1980s cars.

I don't doubt that the last few years have seen significant and rapid progress in AI. What I mistrust is the crescendo of hype, which echoes the tech bubble of the late 1990s and early 2000s. The risk: buying into over-exuberant promises rather than products with a proven return on investment (ROI). Hype clouds our judgment, sometimes intentionally.

If you share my skepticism, but likewise sense an important opportunity and want to avoid excessive caution, here are six questions to help you tune your BS detector.

1. What business problem am I trying to solve?

This is the most important question, and it has nothing whatsoever to do with AI. Granted, a few firms will find value in experimentation, but open-ended projects should be treated with extreme caution. Better to clearly define the business problem you want to solve.

You should evaluate any business investment against three criteria: Will it increase revenue, reduce costs or mitigate risk? Anchoring new technology to at least one of these fundamentals will establish its value. After that, assigning ownership and accountability is the best way to keep a technology initiative on track.

Related: 5 Ways to Make AI Work for Your Organization

2. Why do I need AI to solve this problem?

Maybe you don't. True AI acquires and applies knowledge and skills. It's good for situations in which variability and novelty exist, but it's difficult to build and therefore commands a premium. Consider the complexity of a self-driving car navigating busy city streets. Does your business problem involve continual unpredictability?

Machines that get incrementally better at a task sound compelling, but you need to focus on the outcomes delivered and not the technology used to achieve them. Will you gain a margin of improvement that makes the cost of AI worthwhile? Figure out a test to evaluate the size of the margin. Run it on paper and again in a proof-of-concept project. Make sure AI earns its premium.

3. Do I have sufficient data to use AI?

The best AI solutions outperform people at specific tasks such as recognizing cancer cells in scans or finding errant traders in an investment bank. But teaching a machine to make sense of messy and inconsistent data requires extensive training. AI uses models to make sense of the world and generalize. Finding enough examples to build a good model can be difficult.

Healthcare systems or banks can draw on extensive historic data. Can you? Even if you can, trawling through it to find relevant examples can be time-consuming and costly. To overcome this obstacle, some companies have started to work on AI model-training software that makes the process quicker and cheaper. Even so, it's a tough assignment whenever data is scarce.

Related: Big Data, Big Deal! 5 Ways It's Going to Be Revolutionalized in 2018

4. Should I build or buy an AI solution?

If you're trying to embed AI into your own products or services, an in-house capability might make sense. Still, don't underestimate the resources and specialized expertise involved. A carefully selected partner may offer a quicker path to glory.

If you're tackling a known business problem, you'll likely be better off working with an experienced vendor. Don't be blinded by technology; whatever you buy will need to be customized or adapted to your environment and requirements. Focus on your vendor's understanding of your business domain and the type of data the consultant will need to leverage.

5. How well does the vendor know my domain?

Some vendors claim AI makes domain experience irrelevant. Don't believe it. It's quicker and less stressful to work with a consultant who doesn't need to learn your business from scratch.

Check any potential vendor's relevant experience and partnerships. Can the vendor's leaders give production examples of comparable problems solved for others? If your problem truly is unique, seek experts who bring experience in dealing with parallel challenges -- perhaps in a different industry with similar sorts of data.

Related: 10 Ways AI and Chatbots Reduce Business Risks

6. Is there a proven ROI?

In Gartner's report, the "peak of inflated expectations" is followed by the similarly whimsical "trough of disillusionment." AI will lose its luster as technology buyers look past blithe promises and start demanding proven results.

In my opinion, that's merely sensible business practice. So why be patient? Ask to see ROI measures today. Your smart money is on solutions that don't involve an extended learning curve for the buyer or the vendor.

Related: The Future of Productivity: AI and Machine Learning

While I'm obviously a bit of a cynic, I've seen firsthand the difference that genuine AI can make. Machines can take on tasks that are important but arduous for people -- and do them better. It does more than save time or give a nudge to performance. It opens space for a shift in organizational change that reaps far greater rewards. That's the true promise of AI, but realizing it takes more than clever software.

Amara's law states that "we tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run." It's sound advice. Leaders face the twin challenge of deciding if an investment in AI can create value in the near term while figuring out how to adapt their organizations for a world in which AI is ubiquitous.

Fostering positive experiences and making investments that deliver a healthy ROI will help reveal what's at stake and lay the foundation for deeper change ahead. Heeding business fundamentals rather than hype will allow organizations to make smart choices today and build their experience of AI for the future.

Brett Jackson

CEO of Digital Reasoning

Brett Jackson is the CEO of Digital Reasoning, a global provider of AI-enabled communications analytics solutions to financial services, healthcare and government. Motivated by potential, Brett has over 25 years of experience helping software businesses succeed and grow.

Related Topics

Editor's Pick

Everyone Wants to Get Close to Their Favorite Artist. Here's the Technology Making It a Reality — But Better.
The Highest-Paid, Highest-Profile People in Every Field Know This Communication Strategy
After Early Rejection From Publishers, This Author Self-Published Her Book and Sold More Than 500,000 Copies. Here's How She Did It.
Having Trouble Speaking Up in Meetings? Try This Strategy.
He Names Brands for Amazon, Meta and Forever 21, and Says This Is the Big Blank Space in the Naming Game
Business News

American Airlines Sued After Teen Dies of Heart Attack Onboard Flight to Miami

Kevin Greenridge was traveling from Honduras to Miami on June 4, 2022, on AA Flight 614 when he went into cardiac arrest and became unconscious mid-flight.

Leadership

How to Detect a Liar in Seconds Using Nonverbal Communication

There are many ways to understand if someone is not honest with you. The following signs do not even require words and are all nonverbal queues.

Leadership

Alternatives to Layoffs in Tech: Maintaining a Stable Workforce

Layoffs are not always the best option and can often be detrimental to the organization as a whole. Companies can keep a steady workforce while still controlling expenses and adapting to market changes by thinking about possible alternatives to layoffs.

Money & Finance

How to Choose a Credit Card for Your Startup

When considering the best business credit card for you, take time to weigh the rewards and benefits of each one to determine which card will grow with your business and understand your needs.

Marketing

Entice Customers to Make Additional and Larger Purchases Using These Two Tactics

With transparency, permission, and an eye on the customer's perception of value, you can knock both cross-selling and upselling out of the park.

Leadership

Female CEO Shares Her Experience Leading an AI Company

Kerry Goyette blazed a trail by leveraging her behavioral scientist roots to reimagine a new use for AI.