Applying the Framework to AI Systems: The Good Pt 1
Whenever I show clients Viabl’s AI Evaluation Framework, I often get a few blank looks or casual shrugs. They’ll nod and say, “Sure, that makes sense…but why does it matter?”
Here’s why it matters: it can help you spot the AI ideas that are bound to succeed—and steer clear of the ones that could seriously hurt your business.
In this and subsequent posts, we’ll show how our framework applies to AI systems. By “AI system,” I mean a tool that uses generative AI to handle one or more specific tasks, like Cursor, SceneSnap, or the example below. That’s different from general large language models (LLMs) like ChatGPT or DeepSeek, which can do a wide range of things, and it’s also different from using one of those models for a single purpose like image generation.
Here's a quick refresher on the framework:
Viabl’s framework for evaluation of AI tasks
A Good Example: Salesably’s Conversation Practice Studio
Why It’s a Slam-Dunk
Sales training and onboarding can be tough and time-consuming, and it’s especially challenging in the Senior Living industry. Leads are expensive, and each conversation carries high stakes. Salesably’s Practice Studio addresses this by creating a set of Generative AI “Personas” that act like potential senior living residents. This lets sales counselors practice responding to inquiries, leading discovery conversations, handling objections, and following up with previous leads — all without risking real prospects.
This is one of the smartest uses of AI I’ve seen. To see why, let’s run it through the framework.
The Framework
Can You Share the Data?
Absolutely. In this case, the “data” is just the practice conversations themselves. There’s no confidential or personal information involved.
What Happens If the AI Messes Up?
Honestly, nothing. That’s what makes this use case so great: it’s hard to imagine a scenario where the AI’s mistakes cause real damage. In fact, it’s arguably a feature!
Suppose the AI starts going off-topic and spouts random political views. The worst-case scenario? The salesperson gets extra practice in guiding the conversation back on track. In other words, any “hallucination” becomes a valuable learning moment rather than a bug.
As founder John Greaves puts it, “Our real life prospects say all sorts of things that we have to adjust to, so getting a strange statement from a Persona in the Studio is simply part of the game.”
Since there’s no bad outcome, no need to run through the full framework.
This is a great use of AI.
Salesably and the Practice Studio have met with great enthusiasm and it's not hard to see why: their AI system is built on a smart use of generative AI.
Salesably is now rolling out custom-branded Practice Studios for many senior living portfolios across the United States and Canada, and Vidal Graupera and John Greaves are looking for additional vertical markets to explore. If you’d like to learn more about how this technology can help with training, reach out to Vidal@Salesably.ai or Jgreaves@nextstepsalesmanagement.com.
What’s Next?
In the next post, we’ll walk through another one of my favorite AI products.
And then we'll get to look at some real-life AI disasters and how they could have been avoided.
Got a usage of AI you want us to examine? Comment! I’d love to hear your ideas and help you figure out where AI fits best in your business.