BotMan AI

We forge new code
from legacy systems.

At BotMan AI we build Codikel, the engine for legacy modernization at enterprise scale. Knowledge graphs más generative AI. Between 50% and 90% savings in time and cost on large migrations.

Trusted by

Forus case

16 years projected, delivered in 14 months.

0+Forms migrated
0Batch processes
0Months (vs 16 years)
0+Pipeline steps

Codikel: Modernizing your legacy systems

Migrating legacy can take years

Translating a legacy system leaves you with a legacy system in new syntax.

Traditional tools transliterate code line by line and lose the business logic buried under decades of evolution. The result: an application you have to rewrite again within a few years.

How Codikel solves it

Re-architecture, not translation.

Codikel builds a knowledge graph of your entire system — dependencies, business logic, data models. From that graph it generates a new architecture, not a transliteration of the old code. The same graph powers Barbara, the agent that answers technical and functional questions about your system right from the IDE. Turnkey. Production-ready.

About us

A technology company. Not a consultancy.

BotMan AI develops Codikel, an AI-powered platform for modernizing legacy applications. We don't rewrite code by hand — we built the engine that does it for you.

Mission: do in months what used to take years.

We develop technology that modernizes legacy applications at a speed and scale that were previously impossible. Codikel combines knowledge graphs with generative AI to compress entire migrations into a fraction of the manual time.

Vision: the global platform for legacy modernization.

Become the world's reference platform for legacy application modernization, enabling enterprises and technology partners to run enterprise-scale migrations with our technology.

Fortune 500 leadership + generative AI.

Our team combines Fortune 500 technology leadership with deep specialization in generative AI, knowledge graphs, and modern software architecture.

Team: 20+ specialists across two continents.

More than 20 AI specialists, distributed across two continents, working every day on the same engine.

Founder & CTO

Alejandro D'Andrea

25+ years of technology leadership — Fortune 500, generative AI and legacy modernization.

0+Years in technology leadership
0Boards and steering committees
Fortune 500Executive experience

Current role

  • CEO and CTO at BotMan AI — AI-powered legacy modernization platform.
  • Board member of Talento Digital para Chile.
  • President of the CIO Forum Chile Board.
  • Speaker at Gen AI and digital transformation conferences.

Previous experience

  • Former CIO of Walmart Chile (International).
  • Former GenAI Lead and CTO of core platforms and architecture at Falabella.
  • Former Founder and Director of the Digital Transformation and Innovation area at Cencosud.
  • Former Co-chair of the AI & Data committee at AmCham (Chilean-North American Chamber of Commerce).
  • Former Co-director and professor of the Generative AI Diploma, Universidad de Chile.

Education

  • BSc in Computer Science — Universidad de Buenos Aires.
  • MBA — Instituto Di Tella.

The method

Codikel isn't a translator. It's a hybrid pipeline.

60+ steps where determinism matters. Generative AI where it adds real intelligence. The way to deliver the legacy system complete, verifiable and gap-free — production-ready.

Why alternatives stop at 80%

Status quo

Traditional manual migration

  • Extremely high cost
  • Long delivery timelines (years)
  • High probability of human error
  • Fragmented knowledge of the system

Rule-based

Static rule migrators

  • Accelerated migration process
  • Partial migration (under 80%)
  • No architectural restructuring
  • Only literal code translation
  • Incomplete results

Multi-agent

Multi-agent migration systems

  • Accelerated migration process
  • Roughly 80% guaranteed coverage
  • 20% uncertain due to non-deterministic process
  • Gaps, duplications and hard-to-find errors
  • Semi-static quality standards

Codikel

Codikel Migration Service

  • 50–90% savings on large migrations
  • Total migration — 100% of agreed scope, guaranteed
  • Modern re-architecture with clean APIs
  • Follows the client's development, security and quality standards
  • Automatic dead-code elimination
  • Deterministic pipeline of 60+ AI-powered steps
The missing 20% isn't a detail. It's the difference between a system in production and a migration you have to redo within a few years.
0–0%Savings on large migrations
0%Of agreed scope
0+Deterministic steps

The pipeline

From legacy to production in four phases.

Every migration goes through a 60+ step deterministic process, validated by humans and AI before reaching production.

  1. 01

    Discover

    Codikel analyzes your entire system and builds a knowledge graph — dependencies, business logic, data models. Before touching a single line, it understands the whole domain.

  2. 02

    Re-architect

    From the graph it specifies a modern microservices architecture based on your technology standards. Not transliteration: a new form of the same system.

  3. 03

    Generate

    The pipeline produces code in Spring Boot, React, .NET 8 or the stack you choose. Every step is reproducible, auditable and reviewable. Duplicate-code and dead-code optimization.

  4. 04

    Validate

    Automated validation, human review and final optimization before turnkey delivery. The application reaches production without surprises.

Demo

Watch the engine in action.

A walkthrough of the real pipeline: how Codikel builds the knowledge graph, re-architects the system, and delivers modern code that's production-ready — narrated over real code.

Where it applies

Legacy stacks Codikel can modernize.

If your system runs on any of these technologies, Codikel applies. These aren't past demos — they are the stacks where the pipeline has a clear migration path.

COBOL: 220 billion lines still in production.

220 billion lines of COBOL run the world's most critical systems. The average COBOL developer is 58 years old. The clock is ticking. Codikel offers a pragmatic modernization path to Java Spring Boot.

Delphi: native desktop, integrations stuck.

Delphi systems tend to stay tied to native desktop: per-machine deployment, limited modern integrations, and a developer community that shrinks year over year. Codikel re-architects to Spring Boot + React or .NET 8 without losing the accumulated business logic.

Visual Basic 6: unsupported since 2008.

VB6 has had no support since 2008. Every Windows update is a risk. VB6 developers are practically extinct. Codikel parses and migrates VB6 code to a modern, maintainable stack.

Featured projects

Real migrations, in production.

02

More than 4 million lines of Delphi: ERP migration to .NET 8.

One of the most widely adopted ERPs in Latin America. Over 4 million lines of Delphi XE4 migrating to .NET 8 — all modules, 12-month timeline. The same methodology proven with Forus, at 8× the scale.

03

MFO: intelligent documentation
of legacy Java code.

Documentation of the core system with a knowledge graph that connects every component — explained not just by the code, but by the context of its interactions. Result: new developers shipping to production in 2 weeks.

What they say

Voices from teams that have already migrated.

La decisión de confiar en la IA Generativa fue el diferencial clave. No se trataba de “traducir” código, sino de entender la arquitectura, usar grafos de conocimiento, mapear las dependencias y lógica compleja, y sobre todo, preservando el ADN del negocio.

GR
Gustavo Rivera LunaCDTO · CIO, Forus
Read post →

Habiendo trabajado en migración de datos y aprendido IA generativa para acelerar el desarrollo de SW con [BotMan AI], puedo creer perfectamente esos números. Es notable lo que están haciendo!

MM
Marcelo MirandaIT CIO, Confuturo
Read post →

HP tuvo el privilegio de ser clientes tempranos de [BotMan AI] y doy fe del valor que agregaron a nuestro negocio de servicios.

MM
Martin MontesGeneral Manager, HP Inc Chile
Read post →

Deliverables and governance

What you get, what we ask, how we handle the data.

Codikel isn't a black box. Here are the concrete commitments behind every migration — on the BotMan AI side and on the client side.

Service deliverables.

What you receive when the project closes.

  • Migrated code following the client's technology standards.
  • README + Swagger for every generated API.
  • Unit tests with ≥80% average coverage.
  • Functional test evidence — screenshots, logs and validation videos.
  • Proprietary validation pipeline — hundreds of verified data points on the source code, documented and reproducible by the client's team.
  • Additional deliverables — available under separate quote based on the requested scope.

Client commitments.

What we need from you to get started.

  • Technology standards — the client defines target stack, coding conventions and architectural patterns.
  • Code access — secure access to the source repository and the database during the project.
  • Delivery preferences — release cadence, environments and acceptance criteria.
  • Functional tests — written descriptions, video walkthroughs of the critical flows, or direct validation by the client reporting the issues found.

Security and data.

How we handle the code during the project.

  • Code on controlled infrastructure for the full life of the project, with access restricted to the assigned team.
  • Temporary copies on Microsoft OneDrive — deleted at closing. On-prem variants available at extra cost.
  • LLMs via enterprise cloud only (Azure, AWS, Google, Anthropic). No local LLMs due to quality requirements.
  • Authentication — by default we preserve the existing scheme. Active Directory available as an option.

FAQ

Frequently asked questions.

Other BotMan AI products.

A shared technical culture across distinct enterprise products. Each one solves a specific problem, all powered by generative AI.

Metron: intelligent data catalog.

Metron turns undocumented databases into AI-enriched catalogs. It analyzes metadata, sample data, execution logs and source code to infer field descriptions, business logic, surface hidden relationships and automatically classify sensitive information.

  • Automated descriptions — for every table, field and relationship.
  • Sensitive-data classification — PII, health, biometric, geolocation.
  • Entity-Relationship diagrams — with versionable YAML documentation.
  • Hidden relationships — beyond what foreign keys capture.
  • Barbara agent — embedded queries against the data model.
  • Multi-source support — SQL Server, Oracle, SAP, REST APIs, CSV.

From undocumented databases to living catalogs — without touching a line of code.

Intervus: AI technical interviews by voice.

AI-assisted interview platform that evaluates candidates through real-time, bi-directional voice conversations. The AI interviewer uses the STAR methodology to assess competencies in a structured, scalable way.

  • Real-time voice — candidates speak naturally instead of filling out forms.
  • STAR methodology — Situation, Task, Action, Result.
  • Detailed reports — competency scores, qualitative analysis and full transcripts.
  • Multilingual — Spanish and English with automatic detection.
  • Custom templates — generated from job descriptions.

Scalable interviews without losing qualitative depth.

LuminQA: autonomous AI testing.

End-to-end autonomous QA platform that automates the full testing cycle — from test case generation through execution, evidence capture and self-healing when the interface changes.

  • Test generation — from natural language, source code and video recordings.
  • Self-healing — tests repair themselves when the UI shifts. Zero maintenance.
  • Native integrations — Jira, GitHub, GitLab and Bitbucket.
  • Clean export — to Playwright and Selenium for existing pipelines.
  • Proven outcomes — faster releases, fewer bugs, a scalable team.

60% faster releases · 80% fewer bugs · team capacity.

Is Codikel right for your case?

Modernizing your legacy system is simpler than you think.

Book a free 30-minute discovery session to evaluate whether Codikel fits your system.

Working with a technology partner?

Codikel is also delivered through our network of certified partners. Same platform, same processes — ask your partner if they're in our network, or write to us to confirm.