Rianex

Systems that work today

Marketing agencies, event companies and clinics: the bottleneck they had, what we built and what changed, in numbers. And how we keep the AI under control.

Eventos BarcelonaEvents · Barcelona

Eventos Barcelona

Proposal automation: −85% client response time.

Challenge

Each proposal took 1 to 3 days to go out. By the time it landed, the client had cooled off or already asked a competitor for a quote.

Solution

We created an intake form filled right after the call, which generates the proposal automatically in GoHighLevel, ready to review and send in minutes, with follow-up reminders scheduled in.

Stack we built

  • GoHighLevel

−85%

client response time

Read the full case study
Tribeca MediaMarketing · AI agents

Tribeca Media

Performance and content AI agents: 0 campaign actions without human approval.

Challenge

Their KEYWORKERS platform had an agent that read Meta Ads campaigns but did not act on them: no Google Ads, no creatives and no way to know how much it spent on AI or when it failed.

Solution

We built the Performance Agent, which reviews Meta and Google Ads, proposes pausing, cutting budget or creating campaigns, executes only with human approval, and generates new creatives when an ad fatigues. We are now building the Content Agent.

Stack we built

  • Python
  • FastAPI
  • PostgreSQL
  • Claude
  • Meta Ads
  • Google Ads API
  • GPT Image 2

0

campaign actions without human approval

Read the full case study
Growth4UMarketing · Lead gen

Growth4U

CRM automation: 100% of the sales process automated.

Challenge

Capture was 100% manual, with no qualification or systematic follow-up: leads went cold in spreadsheets and traceability was lost on every opportunity.

Solution

We implemented a capture funnel with automatic scoring and follow-up in GoHighLevel, leaving the entire sales process measured end to end, from first contact to close.

Stack we built

  • GoHighLevel

100%

of the sales process automated

Read the full case study
Hospital CapilarHealthcare · Madrid

Hospital Capilar

With Growth4U · Lead qualification + Booking SDR: −32% cost per patient (CPP) in 8 weeks.

Challenge

They captured plenty of leads but with no upfront filtering or visibility into the funnel: they didn't know which contacts were worth working or where opportunities were being lost.

Solution

With Growth4U, we built a GoHighLevel system that qualifies every lead with a scoring quiz, runs automated WhatsApp follow-up and, with a Booking SDR, books the appointment straight into Koibox. The whole funnel is measured end to end: conversion, drop-off, appointments and sales.

Stack we built

  • GoHighLevel
  • WhatsApp
  • Koibox

−32%

cost per patient (CPP) in 8 weeks

Read the full case study

How we keep the AI under control

Launching an agent is the easy part. The hard part is keeping it doing the right thing as models, tools and data change. This is how we do it:

  • A person approves before it acts

    At Tribeca Media, the agent proposes pausing campaigns or cutting budget on Meta and Google Ads, but runs nothing without a person's sign-off. Every action is logged.

    How to design human approval (in Spanish)
  • Tests before every change

    Tribeca's system has almost 1,000 automated tests. On a customer service agent, before switching model or instructions we run a set of real conversations: if it gets worse, it does not change.

    What to review before scaling (in Spanish)
  • AI cost, measured and capped

    At Tribeca, the cost of every AI call is recorded per client and has a daily cap. The model is chosen by its cost per correct answer, not by the price per token.

    How to measure an agent's cost (in Spanish)