Rafael Petinati · Technology consulting
Technology and AI for lending operations
I was a partner and the person responsible for all technology at a payroll-lending platform — around 60 partner agencies, 29 cities, R$ 78 million originated in 2024. Today I bring automation, AI and regulatory readiness into operations like that one.
The credential
I have run an operation like this from the inside
For fifteen years I was a partner at Grupo MRX, responsible for 100% of its technology: CRM, dialer and telephony, sales pipeline, contract management, marketing and bank integration. The system stopped being internal and became a platform — operated daily by dozens of partner agencies across the state of São Paulo.
The figures refer to origination volume that flowed through the platform — not revenue. Aggregate data; no partner is identified.
What I do
Two ways to work together
Scope, timeline and price agreed before anything starts. No open-ended hours, no scope that grows halfway through.
Operations and AI Assessment
3 weeks · remote
A map of the operation, automation and AI opportunities ranked by return, a read on regulatory readiness, a recommended architecture and one working prototype built on real data. Ends with a 90-day plan and estimated cost.
Technology Leadership
Fractional CTO · 12 months
I own the technology of the operation: roadmap, architecture, vendor selection and management, team leadership, regulatory readiness and execution of the AI agenda. Equivalent to two days a week, remote or hybrid.
And then
Assessments usually lead into delivery projects — auditing 100% of calls with transcription and script checking, a voice agent for triage and qualification, a contract pipeline with audit trail and traceable consent, a unified data layer over existing systems, bank integration. Scope and price come out of the assessment: quoting before looking at the operation is guesswork.
Who hires me
Who calls
They grew; the technology did not keep up
Consolidation left larger operations running on systems built for a different size — now under requirements for audit trails, traceability and integration.
Channel and correspondent network teams
Onboarding, integration, control and auditing of the network that survived the market tightening.
Operations above 30 seats
Script compliance, cost per seat, service quality and auditing of 100% of calls — not a sample.
Product ready, technology without an owner
Companies with a product and customers but no senior person accountable for architecture and technical operations.
How I work
No surprises halfway through
01
It starts with an assessment
No large project begins before I have looked at the operation. It protects both sides: you are not buying a promise, and I am not selling a guess.
02
Fixed scope and price
Timeline, deliverable and value defined before we start. What changes later gets discussed, not billed on the side.
03
The person you talk to is the person who builds
I do not assemble a team to hand the work off. Whoever discusses the architecture with you writes the code.
Remote or hybrid
Based in São Paulo. On-site when it makes a difference, not as a formality.
No reselling
I take no vendor commissions. If I recommend a tool, it is because it fits the operation.
Written down
Architecture, decisions and operations are documented. If I leave, the company keeps running.
Few clients
Two or three at a time. That is what allows going deep into each operation instead of skimming.
Track record
Thirty years, two chapters and the current one
I am Rafael Petinati. Thirty years writing software, fifteen of them accountable for an entire company's technology — including the parts nobody wants: servers going down at 3am, vendors disappearing, a team to lead, a budget to defend in front of the other partners. That part, not just the code, is what goes into the work I take on today.
São Paulo · BrazilFifteen years
Grupo MRX
Partner and owner of all technology at an insurance brokerage that became a lending platform: servers, systems, processes, vendors and team. The company reached 70 employees with its own call center, and the system came to be operated daily by dozens of partner agencies.
Since then
AI in production, not in slides
The last few years went into building end-to-end AI systems — agents, orchestration, software generation. Not reading about the subject: running code. That is the part that now goes into clients' operations.
If your operation is at that point,
it is worth a conversation.
No commitment and no sales deck. Tell me what hurts and I will say whether I can help — including when the answer is no.
Get in touch