Most reps use AI. Few use it well, and the gap is not about tools or plans — it is about habits. Ten specific practices separate the reps producing genuinely better sales work from the ones producing the same work slightly faster.
1. They assign a role before every task
Not "write a cold email" but "you are a senior SDR with ten years in financial services — write a cold email." The role sets expertise, vocabulary, and judgment. Generic prompts produce generic output, every single time.
2. They build persistent projects for major deals
Not one-off conversations that reset. Persistent workspaces with instructions, uploaded documents, and running context on the deal, so every conversation builds on the last rather than starting cold.
3. They use the prospect's words, not their own
After discovery, they pull the three to five most specific phrases the prospect actually used and paste them in verbatim. "We're flying blind on forecast" goes in exactly as spoken — not sanitized into "they have forecast visibility challenges." The prospect's language is the asset; your paraphrase throws it away.
4. They never accept the first draft
The first output is always a draft. One specific iteration — "cut to 100 words," "open with the funding trigger," "make the ask softer" — changes it completely. Reps who send first drafts are the reason people think AI writing is bad.
5. They chain prompts for complex work
Research in one prompt, analysis in a second, writing in a third, each building on the last. Complex outputs require compound workflows, and trying to get all three from a single request produces something mediocre at all three.
6. They save what works
Prompts that produce strong output get saved — not left buried in chat history, but organized by use case. Over ninety days this becomes a personal library that is worth more than any generic tool, because it is tuned to how you actually sell.
7. They match the tool to the task
A live-web search tool for research with citations; a strong reasoning model for analysis and synthesis of what that research found. The best reps know which tool wins which job and stop asking one product to do everything.
8. They verify every number before it reaches a CFO
AI can produce a confident, wrong statistic. Any figure going into a proposal gets checked first, and calculations built from the prospect's own data are always preferred over generic industry stats you cannot defend.
9. They ask for the reasoning, not just the answer
For deal diagnosis or qualification analysis: "think through this step by step before giving your recommendation." Seeing the reasoning lets you catch a bad assumption before it quietly infects the output you act on.
10. They use AI before the conversation, not after
Pre-call research, pre-demo briefs, pre-review pipeline audits. The work happens before the high-stakes conversation, not in a panic afterward. AI as a preparation tool rather than a repair tool is the habit that separates the top group most clearly.
Prompt: "Here is how I currently use AI in a week: [DESCRIBE]. Score me against these ten habits — role assignment, persistent projects, prospect language, iteration, prompt chaining, saving what works, tool matching, verifying numbers, asking for reasoning, and preparing before rather than after. Name the two habits that would improve my output most, and be blunt about which ones I am skipping."
Prompt: "Take the three sales tasks I do most often — [LIST THEM] — and rewrite each of my current prompts to include a role, real context, an explicit format, and one iteration instruction. Show the before and after so I can see what changed."
What unites all ten is that none of them are about the tool. They are about preparation, specificity, and refusing to accept a first draft — which are the same habits that made reps good before AI existed. The technology did not create a brand new skill so much as raise the return on a set of old ones that always mattered.
Frequently Asked Questions
Which habit should I adopt first?
Using the prospect's exact words. It costs nothing, takes seconds, and it is the single fastest way to make output stop sounding generic. Most reps skip it because paraphrasing feels more professional, which is exactly the instinct that flattens their emails.
How long before these habits show results?
Individually, immediately — a role assignment improves the very next output. Collectively, the compounding shows up over a quarter, as the saved prompts and persistent projects accumulate context you no longer have to rebuild.
Is a personal prompt library really worth maintaining?
It is the highest-return habit on this list over time, because it turns every good output into a repeatable one. Reps who rely on memory quietly rebuild the same prompt fifty times over a year; reps who save it start each time from where they last left off.
Put It to Work
Pick the two habits you are skipping and apply them to your next three tasks. Browse the library for prompts built on all ten.