The meeting was supposed to be a typical sales kickoff.
Product updates. Q1 goals. A few slides about where the year was headed.
But somewhere between the morning sessions and lunch, I ended up at the front of the room — not presenting, not pitching. Building.
The account executives had been talking through their biggest frustrations. Researching a new prospect before a call takes too long. Writing personalized outreach is slow and repetitive. After a demo, pulling together follow-up notes feels like starting from scratch every time.
I work as a Staff Product Designer at a Series C construction tech company. We build software for the commercial construction industry — complex contracts, high stakes, fast sales cycles. I’d been living inside these same problems for a while.
So when the team started listing their pain points, I said: “Give me a few hours.”
Building Live
I sat down with direct input from the sales team. What do you need to know before every call? What makes a follow-up email actually get a response? What kills your momentum after a demo?
They told me. I built.
A few hours later, we had a working agent. It could:
- Pull pipeline data from Salesforce and build morning briefings for each rep
- Prep for forecasting calls with deal summaries and risk flags
- Highlight at-risk leads before they go cold — stalled deals, missed follow-ups, aging opportunities
- Draft outreach emails in each rep’s voice, not a generic template
- Build a meeting prep sheet in under a minute — who you’re talking to, where the deal stands, what to push on
- Write follow-up notes after a call and flag clear next steps
- Answer product questions on the spot using the company’s internal documentation
No IT request. No six-month rollout plan. No vendor contract.
A working tool — built that afternoon, used the next morning.
The CEO Noticed
A few days later, he called it out in the company-wide weekly update. Said the team was “working on a new sales agent on the spot with direct input from the AEs.” Said it was exactly the kind of tinkering that turns into something real.
That recognition changed things fast.
We expanded the approach across the company. Customer success got their own agent — trained on customer health data, renewal timelines, and common objections. Product got one for synthesizing feedback from multiple sources. Engineering got one for pulling context across tickets and documentation. Marketing got one for drafting content that actually matched the brand voice.
Five departments. Five tailored agents. Each one trained on that team’s actual work — not a generic AI tool that anyone could have bought off the shelf.
The company eventually added AI agent usage to performance reviews. Not as a bonus metric. As a baseline expectation.
Three Things I Learned
Speed beats planning. The sales team didn’t need a roadmap. They needed something that worked today. Building in the room — with the people who’d actually use it — meant we skipped months of requirements gathering and got it right on the first pass. The best AI tool is the one that gets used.
Trained beats generic. There’s a real difference between a ChatGPT prompt and a properly built AI agent. The tool I built knew our product, our deal history, our customer profiles, our voice. A generic AI tool knows none of that. Generic tools can do general things. A trained agent can do your things.
Adoption happens through demonstration, not documentation. The reps didn’t read a manual. They watched it work in front of them. That’s it. People adopt tools they see working — not tools they’re told are useful.
Why I’m Building This for Small Businesses Now
I didn’t plan for this to become a service. It found me.
After deploying agents across five departments, I kept thinking: this same approach should be available to small businesses. The tools aren’t exclusive to big tech companies. The methodology isn’t complicated. What small businesses are missing is someone who can build it for them — trained on their actual business, not a template that looks like everyone else’s.
An HVAC company in Seguin. A real estate team in Boerne. A custom home builder in Comfort. These businesses have the same repetitive tasks as a Series C startup. They just don’t have an in-house technologist to build them a solution.
That’s the gap I fill.
When I talk about an AI Employee, that’s what I mean. Not a chatbot that answers FAQs. A trained assistant that knows your services, speaks your language, and handles the work your team does manually right now. Follow-ups. Quotes. Customer questions. Meeting prep. Whatever the repetitive stuff is for your business.
I’ve built it at scale, across multiple teams, inside a fast-moving tech company. I know what makes it work and what makes it fail.
Now I build it for the Hill Country.
Let’s Figure Out What Yours Looks Like
If your team is spending hours on tasks that should take minutes, I want to talk.
Not to pitch you. To show you what I built — and then figure out what the right version looks like for your business. Same approach, your context.
It’s a free 30-minute call. No jargon, no proposal you didn’t ask for, no obligation.
— Ryan Johnson
RJ Digital • Seguin, TX
