Syed Ahmed is SVP of Engineering at Act-On Software, a leading marketing automation platform.
AI integration for better coding efficiency is an easy win, but with other departments in the mix (customer success, marketing, HR, etc.), that integration is not always as streamlined. Giving AI the robust support it needs within your whole company requires buy-in from people who might not understand how AI can best benefit them. Although this is something that likely takes more than one meeting, investing time and energy into doing it right will pay off significantly.
Getting everybody on board is one of the first challenges, but I see it as a game of poker. You would rather be playing with a robust, full hand (all departments onboard) rather than a nigh-empty one (departments having a vague understanding of AI’s true potential).
Let’s explore ways to bridge the gap.
Simply gathering an incidental understanding of AI is a thorny issue. If you don’t start by defining terms and formulae, conversations surrounding AI become challenging. When equipped with the proper vocabulary in an AI context, you can begin to have more meaningful conversations about its benefits. This falls on tech leaders to be able to translate and communicate these ideas clearly to other departments and customers.
For example, a predictive scoring model in marketing can use AI to deliver more accurate predictions about the likelihood of a lead converting. That information needs to be presented in a clear, actionable way for marketers. Transposing a dry formula or chart into a thoughtful, value-added experience offers an attractive reason to pay an additional, modest monthly fee for your AI integrations. Statistics can only do so much; one must present AI in a package the customer needs and values.
As far as AI formulas for scoring models go, the classic Matthews correlation coefficient (MCC) has a lot to offer. A 2020 U.K. study published by BMC Genomics stressed how MCC is both “intuitive and straightforward” and works on both balanced and unbalanced datasets. Furthermore, marketing and onboarding teams can use a more accessible explanation of MCC to better sell customers on why a feature such as predictive scoring in relation to MCC is worth pursuing.
Whether we’re talking about AI or a significant feature change, each department has its own perspective about its needs. Different departments can be optimistic (“Think of all of the time savings!”) or pessimistic (“I don’t want to rewrite an AI’s code.”).
Allow time for the salient stakeholders to explain those expectations and needs fully. Although this can take a lot of time upfront, laying everything out on the table first can clarify objectives and provide heightened focus later. If you don’t define terms and explain how things work upfront, it can result in a lot of wasted time, with multiple teams or individuals having wildly different understandings of core concepts.
Don’t be afraid to use historical data to show past trends and how AI integration can retroactively prove the effectiveness of future integration. Per our example above, a predictive lead score can seem like an arbitrary number, but if you can explain the score with a concrete history of data, it becomes proof of actionable KPIs.
AI is more than a fad; it is the future, and it is the now. Notable company-wide progress can be achieved through effective, deliberate communication across multiple departments about AI’s best utility for each team.