AI Startups: Here’s Your Guide to Get Beyond the Hype and Drive a Commercially Successful Business

My LinkedIn and Twitter feeds are flooded with AI news these days, as are most popular news media outlets. And with good reason: McKinsey claimed in a report earlier this year that AI could create between $3.5 trillion and $5.8 trillion in value annually across 19 industries, ranging from healthcare to retail to agriculture and high-tech. AI hype seems to be at an all-time high – even my teenage son talks about taking AI programming language Python rather than Spanish!

Data may indeed be the new oil, and the AI engine can drive tremendous value for companies from all the data, images and videos being generated today. But if AI is that important, why are we not yet seeing next-gen AI companies lift off and create commercially successful businesses the same way cloud computing and cybersecurity companies have?

At Battery Ventures, we have spent the last several years examining the end-to-end AI technology stack, from big-data building blocks to AI modeling tools and, now, vertically oriented AI companies. Based on our body of work, we’ve seen five winning strategies—not all of them intuitive–that AI companies can use with customers to drive commercially successful businesses and create lasting value.

Know your Target Buyer

One truism for AI entrepreneurs is that buyers of AI technology are generally businesspeople steeped in their particular industry, not technology experts. They are searching for business value amidst all the AI-related talk about deep neural nets and unsupervised learning. So when you’re talking to potential customers, cut out the tech talk, speak in plain English and ground your pitch in concrete solutions to your client’s real-world problems.

One example: Fashion merchandisers are constantly thinking about what types of clothing will drive growth in the next season. If you have AI technology that could help solve this problem, show these fashionistas how image-matching technologies can drive the perfect recommendation for shoes to match a shopper’s dress, and in the process drive higher conversion for their brands. I’m betting these merchandisers could not care less if you leverage Tensorflow or the MXNet framework under the covers to drive your technology—they just want to sell more clothes. So focus on that.

Zero in on the killer use case

Remember: you’re selling a painkiller, not a vitamin. Identify an industry’s most excruciating pain point and offer your product as the antidote. Take AI startup Tractable, which helps automate the insurance-claims process through computer vision—and, in the process, helps accelerate customers’ recovery from auto accidents and natural disasters. The technology can be used to help quickly settle some simple insurance cases without the intervention of human appraisers, freeing up the appraisers to work on more-complex cases and saving, potentially, hundreds of millions of dollars.

Tractable isn’t solving every problem claims adjusters and consumers face, but its focused solution addresses crucial issues and offers a real answer to a thorny business problem.

The same can apply to AI startups in other industries. In agriculture, for instance, farmers experience huge pain dealing with fertilizer and irrigation costs, but also see their profit margins squeezed by commodity crop pricing. A startup like Prospera can leverage AI to assess crop health and monitor factors  like irrigation, pollination, fertilization and others—as well as supply and demand–to optimize crop yield for farmers.

Focus on Human Assisted AI as the Entry Point

All the talk about AI displacing millions of jobs is counterproductive and undermines the true impact AI can have on global industries. There are several opportunities for AI to automate repetitive tasks, and augment human productivity – human-assisted AI if you will. For instance, a shortage of talent is hurting many industries as one of the longest economic and employment cycles in recent history continues on. AI systems from some new companies, such as Entelo* and Eightfold, can use data to automate many aspects of the recruiting process, such as sifting through millions of job profiles, detecting signals from job seekers, and identifying passive candidates that may be a perfect match for a job. These companies have the potential to save recruiters hours of sourcing effort, letting them focus on engaging and closing well-matched candidates. (Recent Entelo research shows that recruiters spend, on average, nearly a third of their workweek manually sourcing candidates for a single role.)

Similarly Reflektion*, an AI-platform for e-commerce, can leverage pattern matching to assess what makes customers “click” on specific products, while leaving the more-complex art of product design and promotions to the merchandising experts. These use cases are real-world, drive tangible ROI, and enhance productivity for employees while expanding the economic pie.

Figure out a partner strategy–fast

All major industries – insurance, real estate, medicine, and others – have existing stakeholders who in many ways hold the keys to the kingdom. Partner strategically with them and you can get unparalleled access to large buyers while making your partners’ products even more sticky.

Once you’re seen as a reliable partner whose product is built into the infrastructure of an industry insider, your future products have a leg up on the competition. That’s why establishing a strong relationship upfront is so important. Take for example, the rise of AI chatbots. This technology allows computers to communicate directly with customers to assist with online tasks, leading to tremendous labor savings. This technology has become so advanced that 27 percent of consumers aren’t sure if their last customer-service interaction was with a human or a chatbot—and the technology may actually even drive higher customer satisfaction.

But industries use chatbots in very different ways, so it’s important to have a strategic, focused on-ramp into your target industry. Customer-service chatbots for instance, tend to partner with several call-center technology providers who have access to several Fortune 500 companies. Chatbot startups accelerate their deployment time while the call-center vendor can provide a better end-to-end solution by integrating chat-bots as the front end to customer service reps.

Structure your Sales Team with Industry Rolodex and Technology Awareness

Finally: As you build your company, you should be looking for sales and marketing executives to hire who understand your tech more than they understand the intricacies of the industry you’re selling into. Why? Because people entrenched in insurance (or radiology, or financial services) are less willing to think about transforming their industry via AI. But more tech-inclined salespeople bring a fresh perspective. By listening closely to industry veterans’ gripes, they can pinpoint ways AI can productively disrupt their industry. Another structure we have seen work well is sales executives with an industry rolodex, paired with sales engineers who understand the AI technology stack. By pairing the two, you can uncover the value of the tech platform, but make it applicable to the industry problem and drive real results.

Corporations are hungry for AI solutions that will reduce resource-intensive tasks and save them money. If your product alleviates a major pain point, and you can explain the concrete ways in which it solves customers’ day-to-day problems, you can partner with an industry behemoth and gain insider status. With that, your company will be well positioned for liftoff.