Right Message, Right Person, Right Time - Designing Your AI-Driven Comms Dashboard

When pitching a story to a reporter, producer, or directly to influencers/customers over social media, what topics, messages, or features should be highlighted over other topics? Should your messaging change over time, as the product launch moves from month one to month four, how much should messaging change? Pealing the union even more, can story pitches be done dynamically with AI on an individualized basis? The reporter/influencer didn’t cover your launch, but s/he just posted on a related topic, how quickly can AI notify you about it and draft a relevant pitch?

We’ve all heard the expression, “We made the best decision possible given the information we had at the time.” Well, what if decision-making information is coming to us every minute or every hour? We’re collecting information from our website, our social feeds, trade media, national media, and our competitors’ sites, for example, and “the best” information is changing hour by hour. In AI speak, if we work with our data scientists to build and refine a “reinforcement learning" model for communications we are creating something that can guide our decision-making about messaging, content, trends, and crisis response in near real-time.

First, some perspective. The most popular AI models are currently based on “structured learning.” That is, we create or use someone else’s AI algorithm (a complex set of instructions) and have that algorithm run through our data. When it runs our known data and gives us back responses we know are correct, the model is considered “trained.”  We can then give the algorithm new unknown data to process and be reasonably certain the responses or recommendations it gives us back are correct. That’s not every detail behind structured learning but that’s the basic idea and it’s 80-90 percent of what’s used today in AI.

Reinforcement learning is different in that the data we process with this algorithm is unstructured. Gathered from different sources and updating all the time. This Reinforcement Learning algorithm processes the data coming into our communications dashboard as often as we instruct it to do so. It uses the data to continually improve the recommendations it feeds back to us. A short video explainer details where it’s used outside the communications field.

For our purposes, talk to your data scientist about a communications dashboard that updates in near-real-time. Help design the information framework for her/him to gather information and how to instruct the AI algorithm to decide which information is most important and you’ll be on your way to creating a sustainable competitive advantage for your company, where the right messages will be getting to the right people, at the right time, through the right communications vehicle.