Image by Editor
# Introduction
Even though large language models (LLMs) are typically used for boxed, archetypal roles like “writing email messages” or “acting as advanced search engines”, they have a lot of hidden potential. It is just a matter of uncovering their hidden potential for creative problem-solving and expanding it into lesser-explored terrains.
If you are keen to discover new examples of such unconventional things to do with LLMs, this article lists and exemplifies seven of them, going far beyond the usual chat interface and conversations.
# 1. Playing Personal Devil’s Advocate for Decisions
Conversational AI systems are meticulously trained to be agreeable with the end user, no matter what — unless they are told otherwise. Next time you need honest guidance for decision-making, instead of seeking validation, ask the AI to systematically rebut and dismantle your ideas when needed, and to test your logic. For instance, see this example prompt:
“Act as a ruthless but logical critic. Review this project proposal and identify the top three hidden risks or logical fallacies I have overlooked.”
# 2. Decrypting Arcane Technical Errors
This use case consists of supplying an LLM with something like a cryptic log file or a messy, raw stack trace, and asking it to turn this “machine-generated ball of frustration” into a natural language, step-by-step manual to repair the issue. A prompt template like this (where you may paste the actual error log, replacing the part between square brackets) could do the job nicely:
“I am getting this obscure system error:
[paste error]Explain exactly which line is failing in plain English and provide the commands to fix it.”
# 3. Navigating Private Contractual and Legal Language
Unsure of what you are about to sign in a rental agreement, and unwilling to spend the energy needed to go through those endless, obscure pages full of clauses? How about running it through an LLM — ideally self-hosted, for privacy reasons — and asking it to spot red flags?
“Analyze this rental agreement. Highlight any unusual termination clauses, hidden fees, or non-standard liability shifts that a layperson might easily miss.”
# 4. Simulating Historical Figures or Expert Personas
This one is about prompting the LLM to mimic the specialized communication style or philosophical framework associated with a historical figure, thereby breaking out of conventional corporate thinking.
“Critique my modern social media strategy as if you were an advertising executive from the 1960s Madison Avenue. Focus heavily on emotional appeal and brand positioning.”
# 5. Automating “Rubber Ducking” for Complex Logic
This is very useful for having the LLM detect and point out missing steps in a complex workflow or intricate logic puzzle. Explain the complex workflow or puzzle to the model in an attempt to check if your mental map is well aligned with reality. Take this example prompt template:
“I am trying to build an automated workflow that triggers based on these three specific conditions:
[list conditions]Where is the logical gap in this sequence?”
# 6. Building a Hyper-Personalized Skills Roadmap
Use this prompt to build a bespoke syllabus that omits what you already know and focuses exclusively on your specific knowledge and skill gaps, along with niche educational aims:
“I already understand basic Python, but I want to learn data visualization. Create a free, 14-day study plan with daily practice exercises focusing only on Matplotlib.”
# 7. Bridging Real-Time Cultural Context
This is very useful in the realm of international relations for deciphering the tone, formality, and cultural etiquette in foreign communications:
“Translate this email from a new international client, but also explain the subtext, the level of formality used, and how I should respectfully format my reply to match their cultural business standards.”
# Wrapping Up
These seven use cases only scratch the surface of what becomes possible when you move beyond treating LLMs as simple question-answering machines.
Whether you are stress-testing your own logic, decoding legal fine print, or bridging cultural divides, the common thread is intentional prompting — giving the model a specific role, a clear constraint, and a concrete goal. The more deliberately you frame your requests, the more these tools reveal themselves to be genuine cognitive partners rather than glorified search engines.
Iván Palomares Carrascosa is a leader, writer, speaker, and adviser in AI, machine learning, deep learning & LLMs. He trains and guides others in harnessing AI in the real world.
