Guide 6 min read

How would I use KerrOS as a salesperson

Every channel in one inbox, an AI briefing across all of them, a CRM that holds it together. Here's what KerrOS looks like in a sales day.

KT

KerrOS Team

A salesperson’s day happens across a dozen apps. Gmail. Slack. WhatsApp. Two phone lines. Google Meet. The CRM. By the time you actually want to know what’s moving on your pipeline, you’re stitching a picture together out of four different places.

KerrOS replaces that stitching. Here’s what it does for a sales day.

One list, not four

Every channel you use lives in one inbox. Both Gmail accounts. The Slack workspace — including the #sales channel and the DMs from your CSM. WhatsApp. Discord. Microsoft Teams. The missed-call log from every phone line you connect. Sorted newest first.

Communication Hub inbox with channels tree expanded — both Gmail accounts, Slack Northgate with #sales and other channels, Discord, Microsoft Teams, WhatsApp. Message list shows Jasmin Hale's proposal reply selected at the bottom. Right pane shows the full email body with three questions about ERP integration, budget signed off on her side, SOW requested by Friday, signed Jasmin Hale, CTO, Ridgeline Systems

A prospect’s reply lands overnight. A customer confirms a meeting on WhatsApp. A teammate flags an inbound question in #sales. A cold thread has been silent for four days. Four different apps in the old world, one list here.

The first ten seconds of the morning tell you where to spend the first hour.

Ask for the read

Before you start replying, ask the AI for a read across every channel.

AI chat session titled "Morning pipeline briefing". The prompt reads "Summarize my pipeline. What moved yesterday, what's stuck, what needs me today?" The AI response begins "Good morning, Mia. I pulled from Gmail, Slack, Phone, WhatsApp, and Calendar. Here's where your pipeline is as of 08:45 today." A Moved-yesterday table lists a proposal reply from Ridgeline (Gmail), a 22-minute call with Ridgeline (Phone), a WhatsApp confirmation from Harbor Creative, and a Slack mention from a teammate about a prospect's CTO asking roadmap questions. Below, a Stuck section begins with Meridian at 4 days silent

“Summarize my pipeline. What moved yesterday, what’s stuck, what needs me today?”

The AI answers with a structured pass across every channel. Before it tells you anything, it tells you where it looked — Gmail, Slack, Phone, WhatsApp, Calendar — so the scope is explicit. Then: what moved (a proposal reply, a meeting confirm, a Slack mention you’d never have seen). What’s stuck (prospects past your usual response norm). What’s on today (calls, commitments due).

It’s not magic. The AI will occasionally be wrong — it doesn’t know a given prospect is at a conference this week, or that you already heard back through a side channel. You correct it in your head and move on. What it catches that you would’ve missed — a teammate’s Slack mention that was never going to make it to your inbox — is what earns the thirty seconds.

Calls read themselves

Every call on every connected phone line is recorded and transcribed as it happens.

Phone view with a 22-minute call selected. Call list shows the call at top (incoming, recording, with a short summary reading "Proposal walkthrough. Budget approved on..."), then a voicemail, then older calls. Right pane shows the call detail: contact Jasmin Hale of Ridgeline Systems, with Call Notes reading "Proposal walkthrough. Budget approved on her side. Q2 go-live, ERP integration required. SOW committed for Friday." A recording playback bar is visible

After a proposal walkthrough, you’d normally spend five minutes afterwards trying to remember what you actually agreed to. The call is already summarized in the call detail — budget signed off, Q2 go-live, ERP integration required, SOW committed for Friday. The product owner the prospect named mid-call is pulled out of the transcript automatically — the kind of detail that usually gets lost in the minute-twenty mark of a twenty-two minute conversation.

You listen during the call, instead of taking notes during the call.

One record per person

Every message that hits KerrOS is also an activity on a person and a company. No extra step.

Person record in the CRM for Sarah Chen, VP of Engineering at Harbor Creative. Left panel lists her channels (email, phone, LinkedIn, Slack), Location, Company, and the Lists she's on — Q1 Outreach Targets, Decision Makers, Active Deals Q2. Right panel shows the Communications tab: a chronological thread of emails including Q1 Integration Timeline, Security Review, Proposal: Multi-tenant Architecture, each tagged Sent or Received

Open a contact and you see every conversation with them — across every channel — in chronological order. The March proposal thread sits next to the Slack DM from last week next to the transcript from last Tuesday’s call. It’s not a list of emails. It’s a list of everything.

Before a meeting, you don’t hunt through Gmail for the last thread, then check Slack for the last DM, then the CRM for the notes. You open the person, and the record is the record. An unanswered question from three weeks ago jumps out. A forgotten commitment surfaces. That becomes the first thing on the agenda.

The loop closes itself

A meeting is just a call that hasn’t happened yet.

Meeting detail for Pelago — pricing intro call, Friday April 17 2026, 2:00–2:30 PM (30 minutes), 3 participants. Google Meet link visible. A recording section is present at the bottom

Today’s 2pm meeting will be recorded and transcribed, the commitments pulled out and attached to the account automatically. Tomorrow morning, when you ask the AI what moved yesterday, today’s 2pm will be one of the rows — summarized, sourced, actionable.

Nothing disappears. Nothing has to be written up after the fact. The machinery that captures tomorrow’s call is the same machinery that captured last Tuesday’s.

Ask it to act

The morning briefing is the shallow end. The same AI that reads across every channel can act on them.

Ask it to draft a reply to a prospect’s three ERP questions. It pulls from the call transcript, the original email, and your product docs — and writes a draft that answers the questions with specifics, not platitudes.

AI chat with a "Draft follow-up to Jasmin" session open. The user prompt reads "Draft my reply to Jasmin — answer her three ERP questions, confirm Q2 go-live, cc Derek, commit SOW by Friday." The AI response opens "Drafted — pulled from email-008 (her questions), the call-013 transcript (yesterday's 22-min call), and your product docs for the ERP Connector module." Below is a full email draft addressed to Jasmin Hale with Derek Park cc'd, subject "Re: Proposal — looking great, a few questions", body with three numbered sections answering bi-directional sync, error handling, and professional services scope

The draft sits until you decide — send it as-is, edit a line, or ask for another angle. Nothing goes out until you do.

The same pattern scales. Ask it to book a follow-up for Friday afternoon — it checks both calendars, drafts the invite, sends it. Ask it to draft the weekly update for your sales lead from what actually moved across your pipeline — you edit, you send. Tell it to nudge you if a named account goes silent past your usual response pattern — it watches the thread even when you’re not looking.

It isn’t a chatbot that answers questions when you ask. It’s an operator that can act on the same data it reads, because the data is all in one place.

The before and the after

The pipeline doesn’t get quieter. The deals don’t close themselves. The Thursday meeting still has to be prepped, the SOW still has to be written, the cold prospects still have to be nudged.

What changes is the first fifteen minutes. You used to open four apps and stitch. Now you ask one question, get back the picture, and the one thing the AI gets wrong you fix in your head before it costs you anything.

That’s what KerrOS looks like for a salesperson.

Tags: kerros sales use-case