Software studio · Private capital

Production AI where the judgment stays yours.

controlthrive is a founder-led software studio for capital-raising teams — and the maker of PA-OS, the operating system for placement agents, now in closed beta. Every claim carries a receipt, and nothing ships without human review.

Preview of PA-OS: agents handle sourcing, standardization, matching, outreach, and reporting — each one proposes a change with cited evidence that a person approves before anything is written.

In production at
BBVAKPMGPwCFintonicMinsait
01 Software studio

The five jobs of a raise, handled.

We design and run the systems behind the work that decides a raise — your judgment, encoded in software that compounds. Founder-led, and you own everything the system produces, exportable in full, anytime.

  • 01

    Search — investor sourcing and enrichment

    Live search agents find and enrich investors for a specific mandate, with every claim traced to a source you can open.

    The system works the mandate continuously — coverage stays current instead of decaying the way a bought list does.

  • 02

    Standardization — from messy book to typed records

    CSV exports, transcripts, and meeting notes become a typed, relational investor book through suggest-then-approve proposals.

    The book feeds the CRM you already run instead of replacing it, and every mapping decision you make is remembered.

  • 03

    Matching — mandate to ranked shortlist

    Mandates and deals crossed against your investor book produce qualified shortlists: who to approach, in what order, and why.

    Hard exclusions and do-not-contact rules are deterministic; the ranking reasoning stays visible and reviewable.

  • 04

    Campaigns — personalized outreach, drafts only

    Mailbox-native outreach drafted from the record, personalized per relationship, staged for your review.

    Nothing sends without your approval — compliance-by-construction inside your broker-dealer rails, never around them.

  • 05

    Reporting — the work, made visible

    White-label activity and progress reports where every number traces to the record behind it.

    What justifies and defends the retainer — generated from the same living book the other pillars keep current.

02 How we work
01 Context

Gather the mandate, materials, relationship memory, and constraints that usually live in different places.

02 Ship

Feed context to agents and ship systems that turn messy workflow inputs into usable decision support.

03 Iterate

Refine logic with new context, operator feedback, and evidence so the system keeps producing trusted outputs.

03 Product

PA-OS Closed beta

The operating system for agent-ready placement agents.

Your broker-dealer keeps you compliant. Your CRM keeps names. PA-OS runs the raise — and your relationship memory stays yours, whichever broker-dealer you're on.

  • Import that never asks twice. Drop a CRM export. The agent maps it, dedupes it, and remembers every decision you make.
  • A typed mandate in two minutes. Drop the deck. Every extracted field is quoted back to its source page.
  • Reports every number can defend. White-label client reporting where each figure traces to the record behind it.
04 Selected work
05 Client perspective

“Servando is unusually strong at operating in ambiguity. He can take raw, unfiltered business context, understand the underlying need, and move into execution fast without needing a lot of process, templates, or back-and-forth clarification. What makes him especially valuable is that he combines high agency, real business understanding, and strong AI engineering capability. He quickly grasped the complexity of MVV's business and delivered not only software that addressed it, but also a true 1:1 partnership and technical guidance across every stage of development.”

Misha Vasilchikov Managing Partner · MVV Capital Partners

“Servando has a great knack for data science and for building ML systems that stand up in real-world environments. We worked on recommendation systems together.”

Ed Mattia Head of AI · Outerbounds (acq. by Anaconda)

“Servando developed from scratch a AI-matching solution that formed the core of our data processing capabilities, demonstrating strong technical skills and a structured approach that directly contributed to project success. The solutions lead to 70%+ reduction in duplicates and +1 FTE worth of manual effort eliminated.”

Kamil Kapturkiewicz Founder · Databris

“Servando helped improved our lending algorithms and helped us find high impact issues in our fraud system. His contributions delivered high-impact results that made a real difference to our business outcomes.”

Francisco Llaneza Principal Data Scientist · Fintonic
06 Founder-led
Portrait of Servando Torres, founder of controlthrive

My approach is built on high agency, deep ownership, and a bias toward tangible outcomes. Responsibility means staying with the work until there is a real result: something shipped, improved, clarified, automated, or made useful. I do not believe in vague promises or work that sounds impressive but never becomes value someone can actually feel.

The business comes first. Tools are secondary. What matters is understanding the business context deeply enough to operate almost like an extension of the team: absorbing the domain knowledge, translating it into systems, and sharing progress repeatedly.

The strongest partnerships are long-term ones where both sides are genuinely on the same side: thinking clearly, moving fast, and building something that keeps getting better.

07 Writing

Investor Matching Architecture

Read

A governed search architecture for high-trust deal-investor matching systems: intent contracts, CRM normalization, guardrails, evidence-led ranking, and replayable evaluation.

From Search to LOI

Read

A controlthrive essay on why private capital needs specialized workflow agents around the handoffs from search to LOI.

All writing
08 Next step

If the workflow matters, the software path should feel clear.

Bring the process, bottleneck, or review workflow you want to improve. We can pressure-test the idea together and define the right next build.

Source-linked outputsFounder continuityBuilt for day-to-day use