What are the best AI use cases for sales?

The best AI use cases for sales take work off reps around the customer conversation: account research, meeting notes, CRM updates, follow-up drafts, RFP answers, inbound leads and practice calls. Lumen gets back about four hours a week per seller, and Morgan Stanley saves advisors about 30 minutes per client meeting. In every example, a person still runs the conversation.

Every number links to where it was published.

Which sales teams have measured results from AI?

Lumen, Zubs, Hedda, Rebaba, Morgan Stanley, Microsoft, BuilderBase, Airz, Insider, Salesforce, RevenueWell and Vanta each have a published before-and-after number for one job they handed to AI. The jobs fall into five groups: research, meetings and CRM updates, quotes and RFPs, inbound leads, and rep training.

TeamJob handed to AIWhat changedSource
LumenAccount researchAbout 4 hours a week back per sellerMicrosoft, 2024
ZubsProspect research15 to 20 minutes per lead down to 1 to 2Zubs case study
HeddaProspect research, CRM updates12x faster research, zero manual CRM updatesHedda case study
RebabaMarket mappingMarket coverage from 17% to 50% in under a weekRebaba case study
Morgan StanleyMeeting notes and follow-upsAbout 30 minutes saved per client meetingCelent, 2025
MicrosoftMeeting prep, recaps, CRM updates90 minutes a week saved per sellerMicrosoft, 2024
BuilderBaseMeeting admin and email replies70 customer meetings per person per weekBuilderBase case study
AirzQuotes from call recordings30 to 60 minutes per quote down to about 15Digirise
InsiderRFP answersA 100-question RFP from 4 to 5 hours down to 20 to 30 minutesInventive AI, 2026
SalesforceInbound lead follow-upAn estimated 20 hours a week back per SDRSalesforce, 2025
RevenueWellOutreach to inbound leadsMeeting booking rate from 8% to 20%HubSpot
VantaRep trainingNew BDR ramp from 210 days to 72Hyperbound, 2026

Seven of the twelve teams handed over research or meeting admin, work every rep does every week.

What is the best way to use AI for sales research?

Let AI do the reading before anyone reaches out: annual reports, websites, news, hiring and LinkedIn activity. Research that took a rep most of an hour per account takes a few minutes, and the rep goes into the first conversation already knowing why the account might care.

1. Lumen: account research

Lumen Technologies' sellers used to spend a day or two working through a customer's annual report before a call. With Microsoft's Copilot, Lumen says that now takes "an hour or so". The company estimates sellers get back about four hours a week, which it values at $50 million a year (Microsoft, July 2024).

2. Zubs: prospect research

Zubs builds subscription tools for Shopify merchants, and its CEO researched every merchant by hand before reaching out. That took 15 to 20 minutes per lead. An AI agent now reads the merchant's website, blog posts and LinkedIn activity in 1 to 2 minutes, and the time saved goes into customer conversations (Zubs case study).

3. Hedda: research and CRM updates

Hedda uses AI on both sides of its meetings. Prospect research went from about an hour a day to five minutes. After each meeting the CRM is updated for the team, and nobody has updated it by hand since (Hedda case study).

4. Rebaba: mapping a whole market

Rebaba sells battery systems to large enterprises, where finding the right person can take weeks. Researching by hand, it had covered 17% of its market. Two prompts to an AI agent took that to 50% in less than a week (Rebaba case study).

What is the best way to use AI for meetings and CRM updates?

Let AI take the meeting notes, summarize the call, draft the follow-up email and write the outcome into the CRM. The rep reads the draft, fixes what's wrong and sends it. The work after a sales call looks the same every time, which makes it one of the easiest jobs to hand over.

5. Morgan Stanley: meeting notes and follow-ups

After each client meeting, Morgan Stanley's advisors get notes drafted by AI. They check them, create tasks and a draft email in a click, and send the email themselves. The tool also logs the meeting in their Salesforce CRM. Across 15,000 advisors, it saves an average of 30 minutes per one-hour meeting (Celent, 2025).

6. Microsoft: prep, recaps and CRM updates

Microsoft gave its own sales team Copilot for meeting prep, recaps and CRM updates inside Outlook. Sellers saved 90 minutes a week on average, across 133 weekly users surveyed in November 2023 (Microsoft, July 2024).

7. BuilderBase: meeting admin and replies

BuilderBase's team holds 70 customer meetings per person per week. That's about twice what they could manage if they still did the admin themselves. Replying to a long customer email takes one minute instead of 30 (BuilderBase case study).

What is the best way to use AI for quotes and RFPs?

Let AI write the first draft of quotes, answers to RFPs (requests for proposal) and security questionnaires, from past answers or from a recording of the sales call. A person checks the draft before it goes out. At Insider, a 100-question RFP went from four to five hours to under half an hour.

8. Airz: quotes from the sales call

Airz streams hybrid events for businesses and records its sales negotiations. AI drafts the quote from the recording. A quote that took 30 minutes to an hour by hand now takes about 15, which saves the team around 50 hours a month across 100 quotes (Digirise).

9. Insider: RFP answers

Insider drafts its RFP answers with AI. A 100-question RFP that took four to five hours now takes 20 to 30 minutes (Inventive AI, August 2026).

What is the best way to use AI to qualify leads and book meetings?

Let AI answer inbound leads at any hour, ask the qualifying questions and book a meeting when a lead is ready to talk. Keep a person checking until you trust it. Salesforce's agent hands promising leads to a rep, and RevenueWell's sales development reps (SDRs) reviewed every AI email before letting it send on its own.

10. Salesforce: inbound leads around the clock

Salesforce put its own AI agent on inbound leads. It replies to early-stage leads around the clock, answers common questions and books meetings with the ones ready to talk. Salesforce estimates that frees up about 20 hours a week per SDR (Salesforce, March 2025).

11. RevenueWell: outreach to leads who didn't book

RevenueWell sells software to dental practices. For leads who didn't book a meeting, it swapped its email sequences for HubSpot's AI prospecting agent. The agent started in "review first" mode, with SDRs checking every email before it went out. Between February 2025 and February 2026, the share of those leads booking a meeting went from 8% to 20% (HubSpot).

What is the best way to use AI to train sales reps?

Use AI for role-play before reps take real calls. The AI plays the buyer, raises objections and scores the rep afterward. A new hire can run dozens of practice calls in the first few weeks without a manager sitting in on each one.

12. Vanta: practice calls for new BDRs

Vanta had new business development reps (BDRs) practice with AI role-play before their first live calls. Their ramp time fell from 210 days to 72 (Hyperbound, April 2026).

What do the best AI use cases for sales have in common?

Each team handed AI one job that repeats around the sale and kept the customer conversation with a person. Most of those jobs happen before or after a meeting: research, notes, CRM updates, drafts and practice. Only Salesforce and RevenueWell let AI message leads on its own, and both kept a person close.

If you're weighing AI for outbound, read do AI SDRs work? first.

Frequently asked questions

Time back from research and admin, faster quotes and RFPs, quicker ramp for new reps, and in some cases more booked meetings. Lumen's sellers got back about four hours a week, Insider cut a 100-question RFP to under 30 minutes, Vanta's new reps ramped in 72 days instead of 210, and RevenueWell's booking rate rose from 8% to 20%. Put your own team's numbers into the value calculator.
It depends on the job. In these examples, Microsoft's sellers saved 90 minutes a week, Lumen's about four hours, and Zubs up to 9 hours on prospect research. Salesforce estimates about 20 hours a week per SDR on inbound leads. See where a sales rep's week actually goes.
The research: reading up on accounts, finding the right person and replying to inbound leads quickly. Zubs cut research from 15 to 20 minutes per lead to 1 or 2, and Rebaba went from 17% to 50% market coverage in under a week. Keep a person checking emails before they go out, as RevenueWell's SDRs did at first. More on that in how to make HubSpot an agentic CRM.
Hand AI the research and admin around your meetings: reading up on the account before, then notes, CRM updates and a first draft of the follow-up after. Check anything that goes to the customer before it's sent. See why sales reps don't update the CRM.
Use it where your team loses the most hours outside customer conversations, and for practice calls that would otherwise take a manager's time. Vanta cut new BDR ramp time from 210 days to 72 with AI role-play. Judge it on a number you already track, like ramp time or meetings held per rep. See the sales metrics that matter.
Not in any of these twelve teams. Each one used AI to clear the work around the conversation and kept a person in the conversation. Buyers want to feel chosen, and time from a real person shows them they were. See do AI SDRs work?.

How should a sales team start using AI?

Start with one job. Pick the task your reps repeat most outside customer conversations and time it for a week. Hand that job to AI, keep a person checking anything that reaches a customer, and time it again a month later. If the number hasn't moved, you picked the wrong job or the wrong tool.