⛔ STOP assuming AI as ML/LLM should make decisions.
It should create the insight and information and provide it as an input to the decision.
The decision model then consumes the data and insight as the simple input and runs it through the explicitly modeled decisions.
Benefits of this approach:
🔹You do not rely on AI magic
🔹You have the full control over how decisions are made
🔹The AI agents, similar to human actors follow the same policy and procedure
We get this question recently: Is LLM a replacement of a decision model? As some technologists imagine providing all the documents and data to LLMs and ask questions and gets answered.
Dream come true, no upfront work, no hassle, decision done for you. Right?
💥WRONG!
Let's put aside the fact that even the best LLMs today cannot operate reliably based on provided guidelines and it's not that they do not hallucinate as much, we as humans got used to it.
✨Decisions must be based on policy and procedures driven by rules and calculation. There are guidelines on how to make certain decisions, even human workers should follow them, why AI/LLM is an exception?
Similar to the data, which is good for exploration and sparking your imagination but using data for automation is lazy and wrong. Use LLMs to get better understanding of context, draft a model, and look at the massive data from a different angle. That's what it is for.
💣Decisions require explicit models done and understood by you and execution must follow precisely that model. That's how decision automation is different from what we know as AI today.
➡️Learn how to model decisions at https://lnkd.in/gN6JTn-q
–
Follow “The Uncle of DI” if you’re looking for unfiltered insights into DecisionIntelligence, AI, and DecisionAutomation. Hit the 🔔 on my profile to get notified about my daily posts.
Posted here.
Published July 3rd, 2025 at 07:30 am
