What should they consider as they make that decision?
- Customer desires: What do customers say they want regarding AI? Are there groups of customers with conflicting desires? How can these groups either be prioritized or served customizable solutions?
- Competitor Movement: How are competitors approaching AI? Can provide helpful context on the consequences of deciding one way or another.
- Short-term profit incentive: Given “AI hype”, having any AI solution quickly can attract customers in the short-term. However, other customers may leave. What is the delta?
- Long-term profit incentive: Starting the AI implementation process early can give them a leg up in the long term by building knowledge capital, or it could burden them with tech debt.
- Impact on current product: What might an AI salesbot look like and how would it be implemented in the company’s overall workflow?
- Impact on workforce: Are there employees who would be immediately displaced by this chatbot? What are the company’s obligations to these employees, and how can leadership approach this situation with empathy?
- Inherent risks of AI deployment: Safety and reliability concerns best found and communicated by the company’s technology team.
- Opportunity cost: If they did not invest in AI right now, what else could they invest in that would lead to a better product or increase profitability?
Should they deploy an AI chatbot now or wait?
They should deploy an AI chatbot now. There is a clear value addition as identified by their technology team and through the academic research. However, they should take their time with development and test extensively with design partners before a full release. Right now, it seems like they don’t have a strong understanding of their customers’ desires (and segmentation), which should be the backbone of their AI strategy.
I believe there is a “fast follower” advantage in AI, because the state of the art (SOTA) changes so quickly that first-mover capital investment quickly becomes obsolete as the SOTA progresses. In comparison, being a fast follower allows you to skip massive capital investment with minimal lag/loss, especially when the product is a chatbot that isn’t the most important service that PulsePoint provides. However, being a fast follower requires having an agile team with time and several build-measure-learn loops under their belt. As such, starting quickly enables fast followership in the future.
One aspect of AI implementation that is different from earlier technologies is the speed at which it can be implemented, especially in a solution as simple as a chatbot. The technology can be upgraded quickly, especially transitions to newer LLMs. It is vastly more important to have a clear definition of why AI is being used and how it fits into the overall workflow compared to deploying “advanced AI” for a vague purpose. Once they have this clear definition of why and how, they should get started.
