Every sales blog right now is telling you AI will change everything. Most of what they're selling is slop.
Writing from your own material, not from scratch. AI in sales that reads the emails and calls you've already proven before it drafts a line.
Here's the thing nobody says out loud: a huge chunk of your day goes to admin that isn't actually selling. Meeting notes. CRM updates. Follow-up drafts. Research on tomorrow's prospects. Rewriting the same intro email for the fifth time this week because the last template stopped working. That's where AI earns its keep — not as a magical revenue engine, but as the thing that deletes the busywork between conversations so you can have more of them.
This post is the answer to "how to use ai in sales" that we'd give a friend who just got their first AE role. No keynote slides, no "imagine a future where." Five workflows that work today, what to watch out for, and the honest tradeoffs.
Stop starting from scratch
The single biggest mistake sales reps make with AI is treating it like a blank-page oracle. You open ChatGPT. You type "write me a cold email to a VP of Ops at a 200-person logistics company." You get something generic. You edit it for ten minutes. You wonder if you saved any time at all.
You probably didn't.
AI is useful when it works from your own material — past calls, your own notes, winning emails you've already sent, the way your best customer actually described the problem in the kickoff call. The prompt stops being "write me a cold email" and starts being "write a variation of my top-performing email to this new prospect, referencing what they said in our discovery call last week." Night and day difference in output quality.
This is also where most people's setup breaks down. They end up copy-pasting between ChatGPT and Notion, between a meeting recorder and their CRM, between a PDF reader and a proposal doc. Stop copy-pasting between ChatGPT and Notion. That shuffle is where the promised time savings go to die.
The workflows below all share one property: the AI is reading your stuff, not making stuff up.
5 AI workflows that actually save sales reps time
1. Turn meeting recordings into follow-up drafts
Record the call. Upload the audio file. Get a transcript with speaker labels — so you can tell who said what, not one undifferentiated wall of text. Then ask the agent to draft the follow-up email referencing the three action items the prospect committed to.
The kind of AE who runs this after every demo gets the follow-up out in the hour after the call instead of two days later when the prospect has gone cold. Ten minutes of work (record, upload, ask) replaces forty (scrub through the recording, type notes, draft the email, second-guess the draft). For a deeper look at how upload-based transcription compares to bot-based tools, see our AI meeting note taker breakdown.
2. Research a prospect in 60 seconds from your own notes
If you've been in sales for more than a year, you have a goldmine of prospect research sitting in old call notes, CRM exports, LinkedIn-scraped profiles, and closed-won case studies. Most of it is never looked at again. The full account-context shape — how to keep the per-account thread from slipping — is in How Account Managers Keep Client Context From Slipping.
Drop it all into one searchable place and chat with it. "Has anyone on the team talked to someone at Acme Corp before?" "What was the objection pattern for logistics prospects last quarter?" "Which case study mentions the exact use case this prospect described?" One minute of chat replaces twenty minutes of Slack-asking and Google-searching.
3. Rewrite your outreach from your best-performing templates
Take your five best-performing cold emails from the last quarter — the ones that actually booked meetings. Feed them to the agent as context. Now when you ask for a new variation targeting a different persona, it's writing in your voice, with your hooks, referencing the patterns that actually convert for you. Not the generic LinkedIn-cold-email voice you get from a blank prompt.
Founders doing their own sales benefit most here. You're the voice of the company. Training the AI on emails you've already written beats training it on someone else's generic playbook. If you're a founder wearing every hat — sales, marketing, ops — our guide to AI for small businesses covers how to consolidate those workflows into one workspace.
4. Extract action items from weekly pipeline reviews
If you manage a book of 40 deals and run weekly pipeline reviews, the post-review admin eats your Friday afternoon. Pull all of this week's meeting notes into one prompt: "Give me every action item I committed to this week, grouped by deal, with the date I promised." You get a clean list you can work through Monday morning instead of re-reading ten meeting recaps.
5. Draft proposals from your case-study vault
You've written dozens of proposals. The good ones reuse the same battle-tested sections — discovery summary, proposed approach, success criteria, pricing justification. Don't rewrite them. Chat with your old proposals and say "draft a proposal for this new prospect, reuse the success-criteria section from the Acme deal and the pricing justification from the Globex one." Ninety percent of the work is already done; you're just assembling the right pieces.
Where most sales AI tools fail
Three common traps when you're shopping around for how to use ai in sales:
Plugin fragmentation. You end up with Gong for call recording, Clari for forecasting, Outreach for sequences, and ChatGPT in a browser tab — and none of them talk to each other. The "AI" is spread across four subscriptions, each with its own login, and the insights in one tool never make it to the next. More tools, not less work.
Block-based editors that are slow to edit in bulk. Some of the popular notes-plus-AI tools store your content as structured blocks (a tree of JSON objects) instead of plain text. It looks pretty. It's fine for a single page. But ever tried to mass-update 40 meeting notes at once, or have an AI agent refactor a hundred pages of CRM-export material? Plain markdown beats structured blocks for this by orders of magnitude — same reason code editors are fast at editing code.
Team-first tools that assume a 50-seat contract. A lot of sales AI is priced and designed for sales-ops at a 200-person org. If you're a solo AE, a founder running your own pipeline, or a manager of a small team, those tools give you a 10-minute onboarding call, a CSM you don't want, and a bill you can't justify.
What we use (Docapybara)
We built Docapybara for the individual knowledge worker — including the sales rep running their own pipeline. It's how we'd answer the "how to use ai in sales" question for our own team. A few specifics:
- Chat with your PDFs. Drop in your sales collateral, a competitor one-pager, a prospect's annual report, or that 60-page RFP. Uploaded PDFs get parsed into text the agent can search and quote from. Ask "what does Acme's 10-K say about their supply-chain spend?" and get an answer grounded in the actual document.
- Inline databases inside your notes. Track your pipeline in a live database that sits directly inside a markdown page — alongside your call notes, action items, and prospect research. Not in a separate tab you have to remember to open. A deal page can have the CRM row, the call recording, the follow-up draft, and the case study links all in one scrollable document.
- Built for one rep. Not the sales-ops team, not a 50-seat RevOps rollout — just you and your pipeline. No seat minimums, no admin dashboard, no onboarding call.
- One agent, everything wired in. Search, edit pages, create new ones, query databases, transcribe audio — all working out of the box. You don't assemble it from six half-compatible extensions.
- Markdown-native, agent-first. Your notes stored as plain text so the AI is fast, with everything wired together out of the box — no plugins, no configuration weekend.
Try Docapybara free — import your last five calls and see what the agent does with them.
Common questions
Does this work with Zoom / Google Meet / Teams?
Any of them. Record the call however your platform lets you, drop the audio file into a page, and the transcription runs with speaker labels so you can see who said what.
Can the AI read my PDFs?
Yes. Uploaded PDFs are auto-converted to markdown on upload, so the agent treats them as searchable text instead of opaque files. Works for sales collateral, RFPs, case studies, and prospect financials.
Is my data private?
Your account is single-user, cloud-hosted on our infrastructure. There's no team-shared workspace where a colleague can accidentally stumble into your pipeline notes. For full specifics on where the data lives, check the docs.
Do I need to set up plugins or integrations?
No. One integrated agent, works out of the box. No plugin Saturday, no configuring five extensions to get chat-with-notes, chat-with-PDF, and meeting transcription to coexist.
Can I use my own LLM or bring my own API key?
Not today — we run cloud LLMs for you so there's nothing to configure. Bring-your-own-key is on the roadmap but not shipped.
The point of all this
AI in sales isn't about replacing the rep. The rep is the reason the deal closes. AI is about deleting the 40% of your week that isn't actual selling — the note-taking, the CRM hygiene, the follow-up admin, the prospect research — so the rep gets more time in conversations.
If a post-call workflow used to take you 45 minutes and now takes 5, that's a real chunk of your week back. That chunk is a couple of extra discovery calls, or a full afternoon on a stuck deal, or genuinely logging off on Friday. That's what "how to use ai in sales" actually buys you when the workflow is right.
For the full map of AI workflows by job type — not just sales, but marketing, fundraising, research, and more — see our AI for work hub page.
Stop copy-pasting. Start compounding the work you've already done.