Official Impact Report

Claude Impact Lab
Chile 2026

899 applicants. 48 teams competed. 12 finalists. 3 winners. 2 days that changed the conversation about AI and financial regulation in Latin America.

Chile Fintech Forum 2026Espacio Riesco, SantiagoMay 6-7, 2026

Verified impact metrics

Data extracted directly from the production database. Not from the website. Not from marketing. Real numbers.

899
Applications
48
Teams competed
12
Finalists
3
Winners
183
Active team members
178
Accepted participants
20
Evaluators (judges + mentors)
1,834
Evaluations processed
2,000+
Cumulative attendances (2 days)
250
Awards ceremony audience

Selection funnel

899
899 applications
Individual registrations received
178
178 accepted
437 stand-by | 33 rejected
48
48 teams competing
183 active team members
43
43 teams evaluated
Pre-evaluated by Bendi + human mentors
12
12 finalists
4 per track — in-person Demo Day
3
3 winners
1 per track

Applications per challenge

112
Track 01 — Financial Education
67
Track 02 — Cybersecurity
23
Track 03 — Data Protection

The 3 winning teams

Selected with a binary sub-checks rubric with mandatory evidence. Final score = mentor (40%) + Demo Day judge (60%) + bonus. Scores published live on the Zona IA Stage screen.

#1
Track 01 — Financial Inclusion
Knowledge Builders
86.92
/ 100
#2
Track 02 — Citizen Cybersecurity
Nettoxic
85.96
/ 100
#3
Track 03 — Data Protection
Idesify
82.73
/ 100

Top-tier judges and mentors

20 evaluators total — 10 judges and 10 mentors — leaders in the fintech, AI, regulation and banking ecosystem of Chile.

Judges — Demo Day (10)

  • Jazmin JorqueraVice President, FinteChile
  • Josefina MovilloExecutive Director, FinteChile
  • Jose Manuel PenaHead of AI, Habitat
  • Juan ValderramaCaja Los Andes
  • Francisco ProvosteClay
  • Lucas BascunanMaxxa
  • Alejandro D'andreaBotman AI
  • Gabriel HesSonda
  • Victor AndradeLegal counsel, Impact Lab
  • Felipe PachecoClaude Community Ambassador, Bendita IA / Anthropic

Mentors — Phase 1 (8)

  • Felipe BricenoBanco Falabella
  • Miguel PulgarCMF — Chilean Financial Market Commission
  • Nicolas NuckellActive IT
  • Guillermo BecerraBecerra-Ojeda
  • Cesar CabreraValkimia
  • JaimeEcosistema NET
  • Gabriel HesSonda (dual role: also judge)
  • Felipe PachecoBendita IA / Anthropic (dual role: also judge)

Bendi — The wise owl

AI copilot built on Anthropic's Claude Haiku 4.5. It did not replace the human evaluator — it empowered them. Processed more information in 2 days than a human panel could review in weeks.

440
Automated pre-evaluations
Bendi analyzed 44 teams against 10 binary sub-checks each, generating verdicts with textual evidence and exact location in deliverables.
44
Executive summaries
Complete dossier per team: strengths, areas for improvement, attention flags, and 3 suggested Q&A questions for judges.
1,834
Total evaluations
Combining Bendi pre-scores with final decisions from human mentors and judges. Dual write: v3.3 (binary) + legacy (1-5).
RAG
Living knowledge base
Bendi ingested every question and answer of the event, building a cumulative wisdom system about the Impact Lab proposals.

Conversational chat with evaluators

Mentors and judges consulted Bendi in real time: verifying regulatory claims against the Legal Wiki, comparing evidence between teams, deepening technical aspects. Bendi was not just answering — it was a mentor in itself.

Divergence audit

Every time a mentor marked a sub-check differently from Bendi's suggestion, the system logged the divergence for calibration review. Total transparency between AI and human.

Bendi processed 440 automated analyses for 44 teams, generated 44 executive summaries, participated in 1,834 evaluations, and built a RAG knowledge base of the Chilean fintech ecosystem — all in 48 hours, running on Claude Haiku 4.5.

Legal Wiki — Real regulatory data

Teams worked with real legislation, not fictitious data. The Legal Wiki covered the current Chilean financial regulatory framework:

Law 21.521
Fintech Law — sandbox, open finance
Law 21.719
Data Protection — ARCO rights, DPO
CMF + CSIRT
Supervision, licensing, cybersecurity

56,000+ lines of code. 3 developers. Claude Code.

The complete Impact Lab platform was built from scratch using Claude Code as the primary development tool. Three Chilean developers proved that running a national-scale event with frontier AI tools is possible.

FP
Felipe Pacheco
Chief AI Officer
153 commits
HG
Hugo Guerra
Lead Backend
174 commits
MG
Miguel Gil
Lead Frontend
56 commits
56,230
Lines of source code
53,380
TypeScript lines
2,596
SQL lines
383
Commits
328
TypeScript files
97
API Routes
72
React Components
43
Pages
31
DB Migrations
20
Tests

What we built

Public landing with event, challenges, judges, mentors, sponsors and FAQ
Registration with NLP scoring via Claude Haiku for automatic pre-approval
Per-participant dashboard with status, team, deliverables and score
Private team suite with unique access token per team
Leader picks members — code invitation, role management
Evidence upload — GitHub, system prompts, tools, demos
Live transparent score — every team saw its score updated
Mentors panel + Bendi — AI summary + dynamic rubric + chat
Live judges panel — pitch timer, real-time scoring
Public realtime leaderboard via Supabase with on-screen reveal
Complete admin panel — registrants, teams, engines, rubric, roles
Legal Wiki from public data — CMF, SII, SERNAC, BCN, CSIRT
Bendi copilot — pre-eval + summaries + chat + divergence audit
Live winners reveal — on-screen publication during awards ceremony

Stack: Next.js 14+ (App Router) · strict TypeScript · Supabase (PostgreSQL + Realtime + Auth + Storage) · Claude API · Tailwind CSS · Resend · Vercel

AI Fintech Governance Roundtable 2027

First Public-Private AI Fintech Governance Roundtable in Chile. Closed summit where the guidelines for addressing artificial intelligence in the Chilean fintech industry with a 2027 vision were discussed.

Participants

Leaders

Felipe Pacheco
Claude Community Ambassador

Victor Andrade
Legal counsel

Josefina Movillo
Executive Director, FinteChile

Industry

Maxxa · Sonda · Clay · Active IT · Botman IA · Caja Los Andes · Poweredia

Regulators

Public regulatory sector present
(Law 20.730 — Lobby Law)

Topics discussed and agreed

01

Basic standard for AI use in fintech

The roundtable identified the urgent need to establish a minimum standard for the use of artificial intelligence in the Chilean financial industry — a framework that allows responsible innovation.

02

National MCPs — how to implement them

How to implement Model Context Protocols at the country level, connecting AI models with public regulatory endpoints (CMF, SII, SERNAC) so the industry consumes regulation programmatically and in real time.

03

Government vision on AI

How the Chilean State understands the use of artificial intelligence and what benefits it sees for its own operation — from automated oversight to citizen service with conversational agents.

04

Act with 2027 commitments

An act was signed with governance guidelines and embedded ethics, to be developed with public and private institutions as a model for the Chilean fintech industry.

Sponsors and partners

RoleOrganization
OrganizerFelipe Pacheco · Claude Ambassador Chile
Technical PartnerAnthropic
Industry partnerFinteChile ORG
Sponsor Track 01Clay + Caja Los Andes
Sponsor Track 02Sonda
Sponsor Track 03Active IT
Content sponsorBendita IA
EcosystemBotman IA · Maxxa · Poweredia

Relive the Impact Lab

Two days that changed the conversation about AI and financial regulation in Chile. Here is the record.

Event videos

Recaps, after movies and the live event screen from the Impact Lab.

9:41
Recap Chile Fintech Forum 2026
May 6-7, 2026 · Espacio Riesco, Santiago
VIDEO
VIDEO
9:41
ES
After Movie · Claude Impact Lab
Spanish
VIDEO
VIDEO
9:41
EN
After Movie · Claude Impact Lab
English
VIDEO
VIDEO
Claude Impact Lab 2026

Live event screen

Claude Impact Lab · Chile Fintech Forum 2026

A turning point for AI in Chile and Latin America

What started as an idea — bringing Chilean financial regulation closer to the 5 million Chileans who today cannot read it — became the most serious AI-applied-to-fintech event ever held in the Southern Cone. This is the story, in numbers.

+200
Builders
48
Teams competed
12
Finalists
3
Winners
2,000+
Attendances over 2 days
01

All of Chile arrived in Santiago

899 people applied. Builders traveled from Antofagasta, Valparaiso, Concepcion, Temuco, Punta Arenas. University students asked for permission, professionals left their desks, engineers paid for flights with their savings to be at the Impact Lab. More than 200 builders ended up present at Espacio Riesco.

02

48 hours of code that solves Chile

48 teams competed actively. Thousands of lines of code written in 2 days — code that translates financial contracts for someone opening their first bank account, code that detects scams before a grandmother transfers money, code that protects data of microentrepreneurs from poorly configured AI models. Real solutions for real problems.

03

Iteration without pause: mentors, sponsors, regulators

Teams did not build alone. 2 full days iterating with 10 top-tier mentors (Banco Falabella, CMF, Active IT, Valkimia, Sonda and more), validating with sponsor companies (Clay, Sonda, Active IT, Caja Los Andes), working with real data from Chilean regulators curated in the Legal Wiki: CMF, SII, SERNAC, BCN, CSIRT.

04

The Governance Roundtable marked a historic milestone

For the first time in Chile, private companies, government, regulators, non-profit trade associations, and academia sat at the same table to determine how to improve the AI-applied-to-fintech industry with a 2027 vision.

05

The world watched Chile

We had international visitors who saw, in the front row, how the fintech world is developed from Chile — how a small country builds big solutions with AI. Latin America was watching.

06

Always-full workshops

10+ workshops and content sessions on the Zona IA Stage, with silent-stage headphones. Each session with 150-200 people — always full. More than 2,000 cumulative attendances over the 2 days.

07

A high-impact awards ceremony

In front of 250 people, the 3 winning teams took the stage. The 10 judges were top-tier in Chile in AI, fintech and regulation: FinteChile, Habitat, Caja Los Andes, Clay, Maxxa, Botman AI, Sonda and more. Each team earned the prestige of being recognized by a rubric that came from Anthropic — the same rigor applied to Claude products — supported by Bendi, the wise owl.

08

Prizes to keep building

The 3 winners took home a Mac Mini and a standing desk — tools to keep building. It was not a hackathon trophy. It was infrastructure for the next version of their product to keep moving forward.

09

Made in Chile, by Chileans, for Chileans

The platform. Bendi. The Legal Wiki. The workshops. The Roundtable. The judges. The mentors. Three Chilean developers built a complete production platform with Claude Code — 53,000+ lines of TypeScript, 97 API routes, 31 migrations — from Chile, for Chile.

This set positive precedents
for the AI industry in Chile and LATAM

1
AI applied to regulated sectors can be done with real public data and serious results. Chile proved it.
2
Anthropic's Community Ambassador model can activate entire regions in significant ways. It worked.
3
AI-fintech governance is not just from the US or Europe. Chile built its own seat at the global table, from the south.
4
Industry + government + regulators can work together with commitments toward 2027 when there is a common purpose.
The Claude Impact Lab Chile 2026 was not an event.
It was a turning point.