If Generative AI Feels Repetitive, Here’s How to Fix It
Most people use generative AI the same way every day. Open a chat, explain the context, get a usable answer, move on. The next day, start over.
The output is good enough to keep things moving. That is the trap. Nothing accumulates. Each interaction stands on its own, so the quality of thinking resets instead of building. You are not getting smarter help over time. You are getting the same help, repeated.
I run a diverse portfolio: corporate strategy plus several global technology businesses. On a normal day I move between long-term planning and this-week execution, sometimes in the same hour. So, when a tool makes me re-explain who I am and how I decide every single morning, the cost is real. It is the difference between working with someone who already knows how you operate and briefing a stranger from zero, every time.
What changed for me was not the tool. It was the starting point.
I stopped treating AI as something that answers questions and started treating it as something that should understand how I work. Same capability. Different beginning. Instead of starting from nothing, it starts from context. I call this building an AI Executive Assistant. Not a smarter chatbot, but a working relationship where the value comes from shared context instead of one-off exchanges.
Context is the Whole Game
AI performs as well as the context it has, and no better. Without context, the output defaults to generic. Adding context by hand each time helps, but it does not scale. If you have to paste your priorities into the prompt every morning, you have automated nothing.
The fix is to make context persistent, so it is applied automatically. For that to work, the system needs a clear view of five things:
- What you own — your priorities and your focus.
- How you decide — how you evaluate an idea, and what you reject.
- What good looks like — your standard for output, stated plainly.
- How it should operate — your expectations and your guardrails.
- What should persist — the context and decisions that carry forward.
When these are written down, the shift is immediate. You stop reintroducing yourself. The outputs start to reflect your judgment instead of an average of everyone’s. Your work builds instead of resetting.
The hard part here is not technical. It is clarity. Most of us carry these filters in our heads—how we rank a proposal, what makes us stop reading, what “done” means. We just never write them down. Until you do, the system is left to guess. And it guesses generic.
Write the Document You Never Wrote
I built mine as an operator’s briefing. A plain document that says who I am, what I lead, the strategic themes everything has to map back to and the patterns that make me stop reading. Not a personality test. A working manual.
Some of it is voice and language choices. I prefer “use” over “leverage.” I ask real operational questions: what is the real challenge here, what options do you see, what will you do next—and I expect answers rather than rhetoric. Once that is on the page, anything written on my behalf calibrates against it before the first word.
Most of it is judgment. I want a sized prize before an ambition. Addressable market, basis points of margin, numbers with units. I want a named owner and a date, not “the team.” I want to know whether something is a platform or a one-off project, because that determines whether I am interested at all. Those are not preferences I want to retype. They are standards I want applied every time.
There is a principle underneath this that I apply to the whole business: if we can do something once by hand, we should be asking how to do it a thousand times automatically. Writing down how you work is exactly that move, pointed at your own thinking. Do the hard part once, the clarity, and the system reuses it on every task after.
Start at the Destination
I frame most initiatives by naming the endpoint first and working backward. The same approach works here. Decide what a strong AI partner would actually know about you, then build toward it. You do not need a perfect document. You need a real one.
So, start small. Take whatever “how I work” notes you can write in twenty minutes—your priorities, your decision filters, what good output looks like to you—and give them to your AI tool as the standing context for everything you ask next; make this part of the permeant memory of your AI. Then watch what happens over the following week.
Grounding AI in your own context changes the quality of what you get back—not marginally, but enough to notice. I have seen it in my own work.
Looking for more ways to unlock the value of AI? Explore the TD SYNNEX Cloud Labs immersive AI experiences, where you and your customers can experiment with real-world AI use cases, gain hands-on experience with technologies such as Microsoft Copilot and discover new opportunities to demonstrate and maximize the value of AI across different roles and workflows.
Author’s note: If you’re interested in more perspectives on AI, strategy and the forces reshaping the technology channel, I invite you to subscribe to my Substack and join the conversation.
