A recent #MITSMRChat suggests three best practices for organizations implementing AI.

Our recent Twitter chat exploring AI implementation connected more than 150 people wrestling with tough questions surrounding the technology. What do their organizations use AI for? What are the biggest challenges to implementation? And what lessons can we take away from this conversation? Three clear best practices emerged.

Our readers — those of you on the front lines of these efforts — put it best, so we’ve assembled a roundup of some comments that best encapsulate these issues.

1. Consider the “Why”

“Let’s be honest. Most organizations are focusing their AI efforts on things that matter only to them, hardly things that matter to their customers. Best to find the sweet spot where both interests meet.”

Wim Rampen

As Wim observes, organizations often focus on using AI to streamline their internal processes before they start thinking about what problems artificial intelligence could solve for their customers. Consider using the technology to enhance your company’s existing differentiators, which could provide an opportunity to create new products and services to interest your customers and generate new revenue.

Another chatter adds:

“Many companies look to use AI to optimize decision making ([by adding] transparency) which lowers risk/cost and improves [revenue because] AI can help find opportunities that [are] otherwise overlooked.”


Our survey data shows that, while less-advanced organizations focus most AI initiatives on cost reduction, more-advanced companies see revenue increases from the technology, indicating a shift to more strategic — and hopefully customer-centric — AI deployments.

2. Organize for AI

Three years into our collaborative research with Boston Consulting Group on the adoption of AI, we still see the market struggling to align organizations around AI. As one chatter notes, it’s first about developing a shared understanding:

“The biggest challenge is the misconception about what AI is: There’s no one definition. Many companies are excited about it but don’t have resources to consider if it’s feasible for them, i.e., if they have sufficient data.”


Having the proper infrastructure in place is another prerequisite. AI leaders, whom we call Pioneers, place an emphasis on data management and access, laying the building blocks for AI implementation. (More on Pioneers’ distinctive characteristics can be found in our 2018 report “Artificial Intelligence in Business Gets Real.”)

As another chat participant observes:

“AI is a journey, and seems to always be a chicken and the egg problem. Things need to happen in parallel: people, infrastructure, data, software, business processes, and fair, ethical, moral, and legal uses.”

Michael Kanaan

It’s critical to secure top-down alignment, then establish data governance practices to set your organization up for success.

3. Connect AI Initiatives to Organizational Strategy

Finally, we’re observing a nascent shift whereby organizations now think about AI as a piece of their overall strategy, rather than an add-on to it. One can frame this distinction as having a strategy with AI versus only a strategy for AI.

Part of the challenge stems from a disconnect between leadership and those with AI expertise, as one chatter mentions:

“It’s early for many AI implementations, so [the projects are] not well connected to overall strategy. Teams who understand AI don’t often have a voice in developing [corporate] strategy (and vice versa). The exceptions are all-in tech natives.”

Tracy Allison Altman

It’s also important to secure broad alignment around a shared vision:

“Incorporating AI into business strategy requires that everyone understands it enough, wants it, and is bought into it. As a speaker once said in an IBM conference, we have to build both capability and appetite for AI.”

Ron Villejo

The forthcoming MIT SMR-BCG AI report offers “clear evidence … that Pioneers are adopting a more integrated approach to strategy, talking about strategy with AI as opposed to strategy of AI,” in the words of BCG Henderson Institute’s Martin Reeves. We’ll present empirical evidence that organizations that connect their AI efforts to broader digital transformation initiatives see more impact.

The third annual MIT SMR-BCG artificial intelligence global executive study and research report launches Oct. 15, 2019.

Sign up for our monthly email update below to be notified of the report’s launch.