Closer. Started as a Favour for My Wife. It Got Out of Hand.

This one’s personal.

My wife works in cold call sales. Has done for years. She’s good at it. Properly good, the kind of person who can read a conversation and know exactly when to push and when to back off.

But I kept noticing the same problem. After every call, there’s a scramble. Capture what the prospect said about their budget before the next dial rings. Jump between a CRM, a notepad, a spreadsheet someone built three years ago. Try to remember which prospect mentioned Q3 timelines and which one didn’t.

The selling itself? That’s the skilled bit. Everything around it, the note-taking, the data entry, the post-call admin, is just friction.

The penny drop

I’ve spent over a decade building enterprise systems. Language models, automation pipelines, the lot. And this problem, capturing conversation data, identifying patterns, surfacing insights, is exactly what AI is actually good at. Not the vague “transform your business” nonsense. Real, specific, useful work.

Writing up call notes. Tracking objections. Scoring performance against criteria. Pulling together campaign reports.

Pattern recognition and data capture. That’s it.

So I did what any engineer married to a salesperson would do. I started building something.

Evenings and weekends

Closer. didn’t come out of a product strategy meeting or a market analysis. It came out of my living room, mostly after 9pm, fuelled by tea and mild irritation at how bad the existing tools were.

I’d ask my wife questions while she was trying to watch telly. “When you finish a call, what’s the first thing you wish you had in front of you?” “If the CRM could just do one thing better, what would it be?” “How do you actually track which prospects mentioned what?”

She was patient about it. Mostly.

The first version was rough. Call transcription, speaker identification, some basic analysis. But even that, even the bare minimum, made her eyes light up. She could finish a call and the notes were already there. Not perfect notes, but a proper starting point. Key moments flagged. Objections pulled out automatically.

“I would have killed for this two years ago” was the exact quote, I think.

What it turned into

That rough prototype grew. Call intelligence led to prospect profiles that got richer with every conversation. Prospect profiles led to campaign tracking. Campaign tracking led to performance scoring. And then someone (alright, it was me) said “what if you could just ask it questions in plain English?” and AI Studio happened.

“Show me every call where the prospect mentioned budget concerns.”

“Which reps are improving their close rates this month?”

It just… answers. No report builder. No SQL queries. You ask, it tells you.

The whole thing turned into an end-to-end platform without me really planning it that way. Each piece existed because a real person, usually my wife, sometimes her colleagues, said “I wish I could do X” and I went away and built X.

Where we are now

Closer. is still in active development. I’m not going to pretend it’s finished, because it isn’t. Features are being added, rough edges are being smoothed, and the AI gets sharper every time I throw more data at it.

But the core works. Transcription, intelligence, scoring, the lot.

We’re at the stage where I want real sales teams getting their hands on it. Running demos and offering trials to people who want to help shape what this becomes. I’d rather build it with feedback from the people who’ll actually use it than sit in a room guessing what they need.

If you manage a sales team and the post-call chaos sounds familiar, drop us a line at hello@halfbytemedia.com. I just want to show you what it does and hear what you think.

Because the best feature requests still come from my wife shouting something from the other room while I’m trying to code.

— Jason