Best practices
Five prompt habits that make teams faster
Watch two colleagues use the same AI workspace and you will see one get a usable draft in a minute while the other gives up after three vague replies. The difference is rarely talent. It is a handful of habits, and habits can be standardised across a team.
1. Give the context you would give a new hire
A model knows nothing about your company, your customer, or your last meeting unless you tell it. The single highest-leverage habit is pasting or attaching the actual material: the email thread you are replying to, the spec you are questioning, the data you are analysing. Teams whose workspace connects to their drives and data feeds do this almost for free; attaching the real document beats describing it from memory every time.
2. State the output, not just the topic
“Tell me about our onboarding flow” produces an essay. “List the top five drop-off risks in this onboarding flow as bullet points, each with one suggested fix” produces something you can paste into a ticket. Format, length, and audience are one sentence to specify and double the usefulness of the answer.
3. Iterate in the same conversation
The first answer is a first draft. Weak users start a new chat when it misses; strong users reply with what was wrong: “shorter, drop the jargon, keep the pricing caveat.” The conversation carries your corrections forward, and by the third turn the model has learned the shape of what you want. Starting over throws that away.
4. Match the model to the task
Routine rewrites do not need the most powerful model, and hard reasoning should not run on the cheapest one. In a multi-model workspace like ChatLite the switch is one click, so make it a habit: fast models for mechanical tasks, reasoning models for anything with multiple steps or real stakes, multimodal ones when images or scanned documents are involved. Teams that never switch are either overpaying or underthinking.
5. Know when to escalate beyond chat
Some questions should not be answered from a model’s memory at all. Anything involving current facts, competitors, regulations, or numbers that will be quoted onward deserves web search for live sources, or a deep research run that returns a cited brief. The habit to teach is the trigger: if you would not act on a colleague’s unsourced recollection, do not act on a model’s.
Make them team habits, not tips
None of this works as a tips email that gets read once. Put the five habits in onboarding, share strong conversations internally as examples, and let people see how experienced colleagues actually prompt. The compounding effect across a team of fifty is measured in hours per person per week, and it costs nothing but attention.