Magazine Article | August 1, 2018

Looking Beyond The Hype Of AI Software

Source: Software Executive magazine

By Christopher Day

A Q&A with a CEO who just raised a $6 million Series A to support a machine learning software solution for marketers.

Founded in February 2015, DemandJump has quickly risen through the ranks of impressive AI solutions. In March 2018 the company oversubscribed its Series A, bringing in $6 million from Revolution’s Rise of the Rest Seed Fund, Flyover Capital, Cultivation Capital, and 4G Ventures. Its AI-powered customer acquisition platform for marketers delivers an average ROI north of 200 percent within 60 days. CEO Christopher Day sat down with Software Executive to talk all things machine learning – from the kinds of sales reps you need in your corner to misconceptions in the software world.

SOFTWARE EXECUTIVE MAGAZINE: The buzz around AI often forgets this kind of software comes with complex sales cycles. What are the most important traits for a sales rep to have when it comes to selling an AI-backed software solution? What advice can you offer other software companies about selling on the value of an AI-driven solution?

CHRISTOPHER DAY: We look for sales reps who are comfortable with a “challenger sales” strategy. Many companies talk about deploying AI-powered solutions, but when you get to know them, it is apparent they like to talk about deploying AI, but don’t really do it. Reps need to know how to sell on value and show potential customers case studies. The difficulty with showing case studies powered by true AI is that oftentimes these are not believed by the potential customer. Or potential customers are afraid of the solution because they believe it makes them look bad. AI can do things that humans cannot do, so helping the prospect become comfortable with the fact that it is OK to not have all the answers and that you can learn much better together is really important.

Sales reps need to be comfortable talking about ROI. They should build a day in the life of the customer and tell an ROI-driven story about how deploying the solution can indeed make their lives easier and even lead to bonuses and promotions. Plus reps need to have an understanding and comfort in describing the true business impact versus just selling a widget that performs a simple function. True AI-powered solutions drive a revenue-focused discussion versus just a cost-savings discussion.

SOFTEX: Related to the kinds of people you need to sell AI, what kind of technical skills do you need to develop and support an AI-driven solution?

DAY: To develop a truly AI-powered platform, the company must have a mix of in-house mathematicians, machine learning engineers, and seasoned engineers with experience dealing with massive data sets. We often come across companies touting AI capabilities, but when looking at their teams, you find they are probably using black box algorithms with little understanding of how it actually works, and this can produce undesired results. All AI is not created equal. This means there are many different types of math theories that are best positioned to solve various pain points or business problems. Understanding how to apply these math theories in unique ways is the key to out-positioning other AI tools and delivering outsized results.

SOFTEX: How much of a focal point was AI when you were pitching investors for your Series A round?

DAY: We believe there is customer and investor fatigue with terms such as AI and Big Data. We think AI is a way overused term, and that many companies using it don’t truly understand what it means. Thus we have chosen specifically not to use the term very often. Instead we focus more on machine learning in our discussions and use cases. Use cases tell the story better than anything else.

"We often come across companies touting AI capabilities, but when looking at their teams, you find they are probably using black box algorithms with little understanding of how it actually works."

SOFTEX: DemandJump’s cofounders have had a long history of successful business ventures and advisory roles. How is leading an AI-driven company more challenging?

DAY: Many businesses are afraid of AI-powered solutions. When they see case studies that consistently deliver double and triple results, they either don’t believe them or get nervous that it will make them look bad. There is a significant population that is just fine with the status quo — they are comfortable running out the clock so to speak. It can be a challenge to convince some prospects to truly lean in to deploy something that will make their businesses much more successful.

SOFTEX: What challenges come with trying to price an AI solution? Are there any insights you can share about DemandJump’s pricing strategy that can help other software companies conceptualize how to best charge for an AI solution?

DAY: This is an extremely complicated subject. There are entire studies out there produced by groups such as OpenView and Price Intelligently. At a high level, we think of it as a matter of: “If we could double your marketing performance, would you be willing to pay X?” The difference between success and failure for companies today is the acquisition and reacquisition of customers more efficiently than anybody else. Our pricing boils down to: “If you could have a provable model of that ROI, would you want to move forward?” That is the simple question.

SOFTEX: What is the biggest misconception in the software world about AI?

DAY: There is a misconception that all AI is created equal. Watson is a good example of this. Just because IBM built Watson doesn’t mean it can solve any problem that exists.

SOFTEX: For software companies reading this who have yet to dip their toes into the AI water, what take-home advice do you have? What pitfalls should they be aware of? What misconceptions are in the software market about AI?

DAY: It will cost you twice as much to build, at least, and take twice as long to build as you think it will. You will fail often, and you have to be comfortable with that. You must have machine learning engineers who understand how to take the math and work with engineers who understand how to build infrastructure that will scale with billions of data points. You will have to start with senior people across the entire team.

SOFTEX: DemandJump’s Chief Scientist Tyler Foxworthy was quoted as saying, “In reality, most AI tends to hide in plain sight.” For software companies eager to get in the AI game, what does this mean?

DAY: This means that a platform simply serves up the most powerful actions you should take right now. When you take those actions you see revenue increase dramatically (and at a much lower cost). It just works; all you have to do is take the action.

CHRISTOPHER DAY, CEO and cofounder of DemandJump, is a serial entrepreneur with multiple technology exits to Fortune 500 companies.