By skyforbes Here are 5 practical ways that genuinely improved the quality of responses for me:
1. Start with what you’re really looking for instead of a broad request like: “I’m looking for new running shoes.”
Add the real context: “I run 10–15 km twice a week, I’m flat-footed, I prefer soft cushioning, lightweight shoes, and my budget is €150.”
The answer changes dramatically when AI knows what matters.
2. Share your constraints. Without constraints, you’ll get generic suggestions.
Try things like: “I need something lightweight because I travel a lot.”; “I prefer neutral design — no loud colors.”; “I’m choosing between two models already.”
Constraints = personalization fuel.
3. Tell it what you’ve already tried. It improves iteration and reduces repetition.
Example: “I tried the Nike Pegasus — too firm for me. Ultraboost was too soft and heavy. Looking for something in-between.”
Suddenly recommendations become tailored instead of random.
4. Add your preferences & dealbreakers. Tiny details change everything:
- preferred fit (wide/narrow)
- must-haves (cushioning / weight / breathability)
- style (minimal / sporty / casual) favorite brands or materials you avoid
These shape the why behind the recommendation.
5. Reuse your personal context instead of rewriting it.
I got tired of repeating the same info every time, so now I keep short reusable snippets like: running profile travel style writing tone productivity setup Paste them in when needed — it saves tons of time and makes results far more relevant.
I’m now experimenting with humique, a small browser extension that lets you build a personal profile and inject it into prompts when you choose to (stored 100% locally), but I’d love to learn from others before going too far.
(If you are interested to try, let me know down below or in private chat.)
Curious to learn from you all: How do you handle personal context today? Do you keep personal snippets somewhere? Have you built your own workflow around this?
Would love to steal your best ideas 🙃