AI for Regulated Environments

Is your compliance document review as consistent and reliable as it needs to be?

At FurlanAI, we build AI systems that are reliable, auditable, and governed — helping companies in pharma, legal, and financial services make compliance and document review more consistent, efficient, and traceable, with human oversight built in.

Designed for: Human oversight · Audit traceability · Model governance · EU AI Act readiness

40%
reduction in manual regulatory review
6 mo
from first conversation to production
2 mo
to deliver a secure AI reasoning layer
15+
years building ML and NLP systems

Proof it works

Every engagement starts with understanding your workflow and regulatory context, progresses through a working prototype tested against real requirements, and ends with a system your compliance team can rely on and your auditors can follow. Systems are designed with EU AI Act compliance, human-in-the-loop controls, and full audit traceability in mind.

Pharmaceutical · London

Compliance AI Assistant

Augmented regulatory document review by integrating fragmented document silos into a single AI-powered workflow — replacing a manual, error-prone process with a consistent, auditable system, with compliance teams remaining in control throughout.

→ 40% reduction in manual regulatory review within six months
Crypto Infrastructure · Global

Secure LLM Orchestration Layer

Designed a proof-of-concept LLM orchestration layer enabling AI agents to operate on proprietary verification knowledge while preventing prompt extraction and IP leakage.

→ Proof of concept delivered in two months
Enterprise AI · Prior experience

NLP & AI at Scale

Built AI and NLP systems at Informatica (acquired by Salesforce) and NewsWhip (acquired by Sprout Social), including a natural language interface used by thousands of enterprise users.

→ Enterprise-scale production systems

How an AI-augmented review workflow operates

Every system we build follows this governance-first arc — from regulatory input to auditable evidence record. Human approval is always a required gate, not an optional step.

📥
Regulatory Update
New regulation or policy change ingested
🔍
Document Ingestion
Internal policies, procedures, and templates processed
🤖
AI Classification
Obligations extracted, relevance scored, gaps flagged
🗺️
Policy Mapping
AI findings traced back to source documents
👤
Human Approval
Reviewer validates, overrides, or escalates — required gate
📋
Audit Log
Full evidence trail — every decision recorded and traceable
AI-assisted step
Human approval gate
Audit evidence
Designed for
EU AI Act compliance Human-in-the-loop controls Full audit traceability Model governance Explainable AI recommendations Regulatory change monitoring

How every engagement works

01
Investigate
Assess the workflow, data, and feasibility before writing a line of code
02
Prototype
Build a working system with real data, evaluated against real requirements
03
Deploy
Deliver to production with governance, monitoring, and performance tracking

Companies come to us when —

Compliance or document review has become a manual bottleneck with growing cost and risk.

An AI pilot looked promising but never reached day-to-day operations.

Accuracy and auditability are critical — errors have regulatory or financial consequences.

Operations depend on processing large volumes of documents, reports, or unstructured data.

Bojan Furlan

Bojan Furlan, PhD

Founder & Lead AI Architect · FurlanAI

I investigate emerging AI technologies, build working prototypes, and push them into production — in environments where reliability matters more than benchmark scores. I have done this across global corporations in enterprise cloud data management, pharmaceutical compliance, and social media intelligence.

Over fifteen years across academia, startups, and a $1.3B ARR enterprise, I have learned that the system around it is as important as the AI model.

We work alongside your internal compliance and regulatory experts — combining AI architecture with the domain knowledge your teams already have.

FurlanAI — Founder & Lead AI Architect, 2024–present
Informatica (acquired by Salesforce) — ML Architect, 2019–2024
NewsWhip (acquired by Sprout Social) — Head of Machine Learning, 2015–2018
University of Belgrade — Assistant Professor & PhD, 2008–2015

Let's make your compliance review more consistent

Tell us a little about your situation and we'll come prepared. We'll identify where AI can improve consistency, traceability, and review effort — and where human oversight should remain in the loop. No pitch. No commitment.

linkedin.com/in/bojanfurlan linkedin.com/company/furlan-ai FurlanAI Ltd. · Registered in Ireland: 776711

We'll respond within one business day.

Thank you — we'll be in touch within one business day to arrange your workflow audit.