In a previous post, we talked about how AI unlocks a new possibility to generate business coaching strategies that are incredibly personalized and lead to better business impact beyond simply completing a lesson. But it's not just any AI or a basic call to ChatGPT; it's an AI that has been given anonymized, but useful details on your business's unique challenges. These details can be used to create highly-personalized and dynamic business coaching tools that can nimbly adjust to nearly any unique situation at scale. We tested the impact of this approach through our prototype application for business owners in Thailand.
Warm vs. Cold State
In the realm of AI-generated business health assessments, we have created two differing versions: the "warm-state" and "cold-state" model.
Warm-State Model: This version offers a dynamic and evolving assessment experience. It begins with a short, five-question survey to learn about the business owner's unique situation before proceeding with the business health assessment (BHA). These details give the model highly-useful context from which to refine the questions of the BHA. We intuitively know that a small retail shop has vastly different needs and struggles than a large poultry farm. And the short-introduction creates a "warm state" for the model to incorporate these unique differences into the assessment.
Cold-State Model: Compared to the warm-state model, this approach used a similar, AI-
generated BHA. However, we withheld the onboarding survey, and so the model was effectively blind to the business type and could only learn if the user
volunteered the information during the conversation. This usually results in much more static advice, regardless of the user's unique circumstances or past interactions. It's akin to reading a general business advice column – helpful, but not necessarily tailored to your specific needs.
Why Warm-State Wins the Race
The warm-state model, with its emphasis on personalization, holds a distinct advantage. By starting with even a basic understanding of the audience, this model can offer advice that’s highly-tailored to the unique situation, and can update and grow over time as that user summary changes. As a business owner interacts with the business health assessment multiple times, it can begin to build upon itself in a million different paths depending on the unique situation and conversation history.
Moreover, as our experiments show, when users are presented with outputs from both models (via screenshots), the stark contrast becomes evident. The warm-state model's BHA is actionable, specific, and directly related to the business type. Meanwhile, the cold-state BHA is useful, but noticeably more generic. Our testing data showed this survey to be remarkably stable no matter the business type or situation - more akin to a static book than a dynamic AI.
Warm State Model Business Health Assessment
*Note: This warm-state model was demonstrated in a method with absolute minimum engagement from the user. This was done in an attempt to show the personalization potential that exists even with a somewhat adversarial user. As the user responds and volunteers with more information, the assessment grows substantially in personalization. Also note the contextualized results that interweaves the survey replies with the business details.
Cold State Model Business Health Assessment
*Note: This demo had no information about a business type, but was able to perform a generic business health assessment. This is likely still useful to the business owner, but misses on the full-potential that can be gained within the survey administration and the compilation of meaningful results to the user.
Looking Ahead: The Bright Future of AI-Personalized Business Advice
These insights have profound implications for the future of AI in business assessments. Personalization is not just a buzzword; it's the key to unlocking real value and impact. As we look ahead, our team is excited to explore dynamic user summarization, integrate external context like industry trends and current events, and improving the output to offer clear action steps.
The ultimate vision is to create an AI system so attuned to a business owner’s needs that it not only provides invaluable insights but also motivates them to return, reassess, and continue on their path of growth and success. This creates a perpetually learning system that can dynamically grow and meet the ever-changing needs of a business in personalized ways.
While the world of AI in business health assessments is still unfolding, one thing is clear: the value of personalization is undeniable. By harnessing the power of the warm-state model, we're trying to make ever-improving progress towards our goal of providing a business coach (both AI and human) for any business owner around the world. And at a low cost that is available to all.