Should You Build Or Buy AI? That’s The Wrong Question

Seventy-one percent of people  expect their company’s artificial intelligence investments to increase in the coming years  (via 2018 McKinsey data), and it’s no wonder that many of the world’s largest companies are deep in the midst of investing in AI capabilities. But where exactly should they be investing?

Initially, many of these conversations have centered around the classic consideration: “Do we build or do we buy?” But what’s become apparent for the vast majority of enterprises I’ve worked with while deploying conversational AI solutions is this: AI is unlike any technology to come before it. For all but a very select few organizations, the question shouldn’t be whether to build or buy AI-driven capabilities. The real question should be: “How do we build an AI-capable organization?”

The Build Versus Buy Debate Is Settled

Countless articles have been written in which the authors compare the merits of building one’s own AI or machine learning platform and those of operating off of and customizing an existing AI backbone to suit the needs of an enterprise. For the most part, the conclusions have boiled down to this: Unless your enterprise is an AI company at its core — dedicated to building a team of data scientists, maintaining uptime and scalability with a team of DevOps engineers, ensuring constant compliance oversight from security architects, and ultimately monetizing it as a core offering — then you may be better off operating from AI solutions that were purpose-built for enterprises.

The simple fact is that building AI is hard for most people, and I believe those who do it well are a part of a very exclusive club. Not only must a team have deep expertise in the core mechanisms of the technology, but it must also be able to develop across stacks that deliver all the necessary security and scalability that today’s enterprises require. That’s no simple task.

All the same, there are still plenty of companies out there drawn to the idea of owning and controlling the internal development of AI projects. Capital One, for example, developed its Eno natural language processing (NLP) tech in-house. Ultimately it will be up to Capital One’s shareholders to determine if this development was a good investment. Certainly, from what I’ve seen, the stock’s movement to date hasn’t proven that it’s a game-changer.

Meanwhile, other companies have looked to bring AI expertise in-house via acquisitions. American Express, for example, acquired Mezi in 2018. Mezi is a travel assistant app that uses AI to personalize the online travel discovery and booking experience. While the complement to American Express’s own travel counselor offerings is apparent, companies risk falling short on an important level when they acquire such solutions. This strategy brings AI capabilities under a company’s own roof, but it doesn’t necessarily instill a deeper AI-capable culture throughout the enterprise.

How To Build An AI-Capable Organization

Becoming an AI-capable organization does not mean redefining your company’s identity around AI technology. It means intelligently infusing expertise into certain areas of your company, empowering it to seek and adopt the right AI-driven solutions for the company’s needs, and then instilling an understanding of the value and capabilities of a company’s AI applications throughout the organization. Here are a few core components of the AI-capable organization:

Centers Of Excellence: While AI-capable organizations don’t need the expertise to design and build full-blown AI platforms in-house, they should maintain a core team of experts who are able to tailor AI solutions to the needs of the company. This is an approach I’ve seen from companies such as Citizens Financial Group and Union Bankshares, both of which appear committed to infusing the right AI integrations throughout their organizations. Moreover, as a recent HBR article explained, companies can create AI oversight bodies with a charter of building an AI vision, identifying business-driven use cases, developing a network of champions, and highlighting success stories.

Managing Training Data: In addition, the AI-capable organization gets a handle on its data for the purposes of continually improving its different AI models. Companies should look to build internal expertise in AI analytics and establish proper flows that can be used to feed training data sets. For example, a company employing conversational AI solutions will want to ensure it regularly assesses logs of customer inquiries to determine the most common ways its unique customers pose questions.

Focusing On Customer Experience: Above all, I believe a company must infuse its teams — from tech departments to customer service units — with a deeply customer-centric mentality. At the end of the day, all AI implementations should share a common goal: contributing to outstanding customer experiences.

After all, what we’re talking about here is becoming AI-capable. Notice that I don’t say “AI-centric.” While many executives might see the appeal in becoming more “AI-centric” in order to showcase the innovative nature of their brand, I’d encourage these individuals to take a step back to ensure their priorities are in order. Technology is not typically an end goal in itself. It’s a means to an end. Given the scarcity of talent in the AI space, enterprises’ competitive differentiators likely won’t come from how big of a technology infrastructure they build, but rather how focused and aligned their efforts are in leveraging AI’s capabilities toward business objectives and serving the customer.

Originally published in Forbes.