AI can read a recorded sales call, or the rep's notes, and propose the CRM updates a rep types today: next step, stage, contacts, competitors and open questions. It also drafts the follow-up email. The rep reviews each change in Salesforce or HubSpot and approves it before anything is saved or sent.
#Key takeaways
- The work to automate is the typing after the call, not the selling.
- AI proposes field updates and a follow-up draft; the rep approves before anything is saved or sent.
- Native CRM AI features suit teams that live inside one CRM; a custom workflow suits teams with their own fields, languages or systems.
- Sales managers get cleaner pipeline data because updates happen while the call is fresh.
- Every proposed change, approval and edit is logged.
#What does the workflow look like after a call?
- The call recording or the rep's notes arrive from the meeting tool or the phone system.
- AI writes a short summary and proposes updates to named CRM fields: next step, stage, decision makers, competitors mentioned, objections.
- AI drafts the follow-up email in the language of the call, with the agreed next step.
- The rep sees the proposed changes next to the current record, edits what is wrong and approves.
- Only then are the fields saved and the email sent, and the approval is logged.
This is a predictable workflow, not an autonomous agent. Anthropic, december 2024 recommends starting with the simplest approach that works. The same pattern drives meeting notes that become tracked action items.
#Why does the rep approve before anything is saved?
Because a wrong stage or a wrong close reason distorts the forecast, and a wrong promise in a follow-up email is hard to take back. The rep was on the call and is the person who knows what was agreed. Approval takes a glance when the draft is right and a correction when it is not.
Approval also keeps the CRM trustworthy. If reps stop trusting what the AI writes, they stop reading it, and the pipeline fills with confident nonsense.
#Should we use Salesforce Einstein, HubSpot's own AI or a custom build?
If your team works entirely inside one CRM with standard fields, start with the vendor's own features. See Stennir vs Salesforce Einstein for where each fits. A custom workflow makes sense when calls happen in Arabic and English, when your fields and rules are your own, or when data must stay in a specific region.
Either way, check what reps already do. The Microsoft and LinkedIn Work Trend Index, may 2024 found that 78% of AI users bring their own AI tools to work. If reps paste call notes into a personal ChatGPT today, a governed workflow inside your CRM is the safer option.
We are
Stennir builds one post-call workflow into your Salesforce or HubSpot, with the rep approving every field and email, and trains the sales team to use it.
We aren't
Stennir is not a CRM vendor, a call recording product, or an agent that updates your pipeline without a rep reading it.
#What happens to deals that go quiet after the call?
The same workflow can flag deals where the agreed next step has passed with no activity, and draft a follow-up for the rep. The rep still decides whether to send it. The worked example of AI follow-up for stalled deals walks through it step by step.
Sales managers can review a sample of approved updates and drafts. The edits reps make show where the prompts, the field rules or the summary format need work.
#What should a head of sales or revenue operations do next?
Write down the fields reps must update after every call, the follow-up they send, and who checks pipeline quality. AI implementation at Stennir puts one workflow into production with real users and a named owner; you pay for finished work. See AI for sales, take the free readiness assessment or book a 30-minute call.
#Common questions
Do you need access to our call recordings?
Only to the calls in scope, and only with your consent rules in place. Stennir agrees with you which recordings or notes the workflow reads, where they are processed and how long they are kept. Many teams start with rep-written notes, which avoids recording questions, and add recordings once the process is trusted.
Does it work for calls in Arabic?
Yes. Arabic and English are fully supported, including calls that switch between the two. The summary and CRM fields can be written in English for reporting while the follow-up email is drafted in the customer's language. Stennir tests on your own call samples before reps rely on it.
What stays with us?
Your CRM, your accounts, your data and the off switch. Stennir builds the workflow into your Salesforce or HubSpot instance with permissions you control, logs every proposed change and approval, and hands over the documentation, so your team can run it, change it or switch it off without us.
What could go wrong?
A summary that misses a commitment, a field update that sets the wrong stage, or a follow-up that promises too much. The rep's approval catches most of these, the log shows the rest, and sales managers review a sample of approved updates to see where the drafts need work.
