Sales leadership · Essay

How I use AI as a sales leader

I use Claude every day to research accounts, prep calls and keep my pipeline honest. It doesn't replace judgment. It gives back the hours that used to go to busywork. Here's the actual workflow.

Brandon Hillier · 5 min read

Most talk about AI in sales is about writing more email. That's the least useful thing it does. The value is in preparation and follow-through: the work that separates a good meeting from a wasted one, and the first thing sellers skip when the calendar fills up.

At Upside, AI was built into how we worked, down to asking questions of our own data in plain English instead of waiting in an analyst's queue. Today I run my own workflow on Claude, with Gemini and ChatGPT alongside it. This is what a normal week looks like.

A week in the workflow

Every morning

A signal sweep on my accounts

A scheduled agent checks my target accounts for leadership changes, funding, new job postings and public filings, then sends me the few that matter. A new revenue leader or a burst of sales hiring is a reason to call, and I hear about it the day it happens.

“Flag new revenue leaders, funding and sales hiring at these accounts, and tell me why each one matters.”

Before every call

A one-page brief in ten minutes

Who's in the room, what's changed at the company, how they make money, the likely objections and the three questions I need answered. I used to skip half of this on a busy day. Now I don't have to.

“Build a brief on this company and these attendees: recent news, how they make money, likely objections and the three questions I should ask.”

After every call

The debrief, while it's fresh

I give it the call transcript and get back the commitments, the open questions, the next step and a draft follow-up. I edit the follow-up and send it the same day, while the other side still remembers the conversation.

“From this transcript, list what each side committed to, the open questions and the next step, and draft a short follow-up in my voice.”

Every week

An honest pipeline review

I ask it to challenge my own deals: which have no dated next step, which depend on a single contact, which haven't moved in two weeks. It's the review a good manager would run, available any time I want it.

“Review these deals. Which have no dated next step, only one contact, or no movement in 14 days?”

Before anything goes out

A first draft, never a final one

Memos, outreach, account plans: AI gets me to a first draft fast. Every fact gets checked against its source, and every word I send is mine.

Where I draw the line

The AI does

Research and account briefs

Call summaries and next steps

First drafts of anything written

Spotting stalled deals in the pipeline

I do

Decide what matters and what to do

Build and keep the relationships

Check every number before a client sees it

Keep confidential data out of tools it doesn't belong in

What it changes for a team

Standard prompts belong in the playbook. When every rep preps a call the same way, coaching gets easier. I can see what they researched and what they missed, and fix it before the meeting instead of after.

The time saved should go to selling. If AI just produces more email, nothing improves. I'd measure it in meetings held and deals advanced, not messages sent.

The rules come first. Which data can go into which tools, and that a person checks every fact before it leaves the building. A team moves faster when nobody has to guess.

My stack: Claude, Gemini, ChatGPT, Salesforce, Gong, Outreach, Glean and Hex.

Building AI into how your sales team works? I'd welcome 20 minutes to compare notes.

Next post: Building a sales org from zero

Email me LinkedIn