I built my own CRM. Then I gave my sales agent the keys.

Instead of renting a SaaS CRM and bending my business around it, I built LNL CRM Command Center: a production system, live since July 2026, where the CRM hands Matteo, my outbound AI sales agent, his daily calling list and he writes every outcome back, email outreach runs on automated sequences, and forecasting, subscriptions, and opportunity alerts live in one place.

ONE CALLING DAY · CRM + agent, both directions ● representative flow

CRMautomation builds today's calling inventory, leads queued for Matteo generated

MATTEOworks the list, outbound calls to clinics calling

MATTEOwrites back after each call: lead status updated, outcome notated on the record updated

MATTEOclinic says yes: demo booked through the CRM booked

MATTEOclinic wants to read first: Matteo sends the VozClinic one-pager from his own WhatsApp or email sent

CRMpipeline and forecast reflect the day's calls updated

The loop runs both directions: the CRM decides who gets called, and no call ends as a sticky note. Every outcome is structured data by the time Matteo hangs up.

20 hrsback per week, between email outreach and cold calling

Jul 2026in production since, running my companies daily

3lines of business forecasted in one console

What's inside

Six modules, one system.

This is the operating console for a multi-brand business, built to spec instead of assembled from subscriptions.

AI agent integration

THE CRM DIALS, THE AGENT WRITES BACK

A daily automation builds Matteo's calling inventory. He works the list, then writes back to the CRM after every call: lead status updated, outcome notated, demos booked, or the VozClinic one-pager sent from his own WhatsApp number or email address when a clinic wants to read before they commit. Yes, the agent has his own email. The agent works the phones; the CRM keeps the truth.

Automated email outreach

SEQUENCES, NOT BLASTS

Multi-touch outreach sequences run automatically against CRM contacts, bilingual where the prospect calls for it, with every send tracked on the record it belongs to.

Financial forecasting by line of business

LNL GROUP · VOZCLINIC · SOL

Revenue and pipeline broken out per brand, wired into live Stripe billing. One screen answers the question every multi-brand founder asks: which line is actually carrying the business?

Opportunity alert feed

SIGNAL, NOT DOOMSCROLL

A news feed tuned to the topics and markets I care about, surfacing opportunities worth acting on inside the same console where I'd act on them.

Subscription tracking

RECURRING REVENUE, WATCHED

Every client subscription tracked in one view, tied to the live billing data, so renewals and revenue changes surface before they become surprises.

Booking page

LNLCRM.COM

Prospect-facing scheduling built into the CRM itself, so a booked call is a CRM record from the second it exists. This is the page Matteo books demos through and my cold outreach closes to.

Inside the system

Real screens, redacted where it counts.

Production screenshots from the live system. Prospect names and contact details are pixelated, and the financial views run on clearly labeled sample data. Real prospects and real books stay private; the software is exactly what I work in daily.

CRM lead record showing Matteo's call note and activity timeline: voicemail logged in Spanish, stage changed from New Lead to Contacted, outreach emails tracked — **The writeback, on the record.** Matteo's own call note ("Left standard voicemail script in Spanish"), the stage he moved the lead to, and every outreach email tracked on the same timeline. This is what "no call ends as a sticky note" looks like.

Pipeline kanban board: 3,199 contacts across 10 stages from No Stage and New Lead through Proposal Sent, every prospect card blurred — **The board Matteo works against.** 3,199 contacts across 10 stages, dragged forward by outcomes, not by me. Every card blurred on purpose: those are real prospects, and this page is public.

Operations console financial dashboard filtered by line of business: revenue, expenses, net and burn-rate cards, a 12-month revenue versus expenses chart, and an expense breakdown donut. Populated with sample data. — **Financials by line of business.** Revenue, expenses, burn, a 12-month trend, and the expense mix, sliceable per brand from the pills up top. Figures shown are sample data loaded for this walkthrough; the real books stay private.

P&L report: expense and revenue categories crossed with each line of business, with per-LOB and grand totals and an Export CSV button. Sample data. — **The P&L, by brand, one table.** Client services, product revenue, and every cost category crossed with each line of business, exportable to CSV. This answers "which brand carries the business" in one look. Sample data shown.

CRM dashboard: 3,199 total contacts and full pipeline breakdown by stage — **The pipeline at a glance.** 3,199 contacts under management, staged from New Lead through Closed Won. Every stage change in that funnel was written by the automation or the agent, not typed by me.

Bilingual outreach template editor with merge variables, spintax, and live preview of a Spanish cold email inviting doctors to call the Sofia demo line — **Outreach that dares the prospect to test it.** A Spanish cold-email template with merge variables and live preview. The pitch: call the demo line, interrupt her mid-sentence, switch to English without warning, then draw your own conclusion.

Full CRM contact schema: custom fields for language preference, PMS system, cross-border patient percentage, WhatsApp status, pilot dates, and more — **A schema built for this exact business.** Language preference, PMS system, cross-border patient share, WhatsApp deliverability, pilot dates. Fields no off-the-shelf CRM ships, because they're mine.

Opportunity alert feed: scored industry news cards on local SEO, AI agents, and clinic markets with save and mark-read actions — **The signal feed.** Industry news scored against the topics I sell into, surfaced inside the console where I can act on it. Signal, not doomscroll.

Booking pages list: six live scheduling pages by line of business, including VozClinic Demo with five bookings and Discovery Call — **Six booking pages, one system.** Demos, discovery calls, and audits, each tied to a line of business. This is where Matteo's yeses land.

Subscription tracking view: monthly and annualized totals with nine active subscriptions, each row tagged DETECTED by the recurring-charge scanner. Sample data. — **Subscriptions, self-discovered.** Every row tagged DETECTED: the system found these recurring charges by scanning transactions, then each one was assigned to its line of business. The chips up top total the burn per brand. Sample data shown.

90-day cash forecast chart with confidence bands, money in and out events, net per day, and projected cash line. Sample data. — **90 days forward, with confidence bands.** The forecast engine projects cash from recurring entries per line of business, each carrying its own confidence level. Sample data shown; the mechanics are production.

How it's built

Production software, not a prototype.

Documented, tested, and operated the way 20 years of enterprise delivery taught me: a written PRD, a schema doc, and phase plans before the first line of code.

FrontendNext.js App Router, TypeScript, shadcn/ui, Recharts for the forecasting views

DataPrisma ORM, NextAuth for auth, live Stripe integration for billing actuals

AutomationResend for outreach delivery, pg-boss for job queues, Anthropic API for the intelligence layer

InfrastructureSelf-hosted on a VPS I operate: Nginx and PM2 on bare metal, deployed from GitHub

In productionSince July 2026, running the real pipeline, billing, and bookings for my own companies daily

Why this matters for you

Your business isn't shaped like a SaaS template either.

Every module in this system exists because my business needed it, not because a vendor's pricing tier included it. That's the difference between configuring software and building it. If your operation needs a tool that doesn't exist off the shelf, this is what it looks like when I build it: spec'd, documented, shipped, and run in production.

It also completes the loop with my other systems: the voice agents feed it bookings and the content engine feeds it attention.