I advise companies in Colombia, Ecuador, Peru and Central America that want to put artificial intelligence to work in their operations: what to do first, with which tool, and how to measure whether it worked.
I'm Carlos Betancur Gálvez, founder of btodigital and a professor at UPB and Universidad EIA. I don't start from a model: I pick the one that fits your case.
Carlos Betancur Gálvez is an artificial intelligence consultant based in Medellín, Colombia. He advises mid-sized companies, executive teams, boards of directors, and marketing and sales teams that want to use artificial intelligence in their operations with measurable results. He works in four phases: assessment, proof of concept, production implementation and continuous improvement, and every project is quoted after the assessment. He is model-neutral: he chooses between Claude, ChatGPT, Gemini or open models based on each company's problem, costs and data. From Medellín he works with companies in Colombia, Ecuador, Peru and Central America, remotely and on site. He is the founder and CEO of btodigital, a digital marketing and artificial intelligence agency, and with his team he builds what he recommends: AI agents, automations and internal platforms. He brings more than 20 years of experience and more than 500 companies advised across 11 countries, and he teaches at Universidad Pontificia Bolivariana and Universidad EIA.
An artificial intelligence consultant helps a company decide where AI delivers a result and where it doesn't, and stays with it until the solution runs on the business's real data and systems. The work doesn't start with the tool: it starts with the process that hurts, what it costs to leave it as it is, and the number that should move.
In practice, this is what I do:
I work with the tools that are running in production today: Claude, ChatGPT, Gemini and open models, almost always in combination. If your company has already decided to implement Claude, that specialty has its own page: Claude (Anthropic) consulting and implementation →
Four phases, with measurable deliverables and the option to stop at any of them if reality doesn't match expectations. Every project is quoted by scope, after the assessment.
I interview leadership and the team, identify the real problem (not the symptom) and assess whether artificial intelligence is the right solution or whether there's a simpler fix. I deliver a document with three possible paths: ambitious, balanced and conservative. If you decide not to move forward, the assessment is yours to keep.
My team and I build a working, scoped version, test it with a pilot group, and measure quality, cost and adoption. This is where ideas that don't work get dropped quickly, before anyone spends months on something nobody will use.
We take the pilot to a stable system: integration with the company's systems, access controls, alerts, error handling and predictable costs. Everything is documented so your team can maintain it and grow it.
Artificial intelligence isn't a project that ends: it's a system you tune. We review metrics, adjust instructions and costs, and add capabilities. It can be a monthly engagement or one-off sessions. You set the pace.
The first step is a conversation.A free 30-minute session over video call or WhatsApp.
Book the assessmentFor the people who have to make decisions about AI and answer for the result. Scope follows the problem, not the size of the company.
They have processes that already hurt and no AI team of their own. We start with one scoped case and a proof of concept, before committing a large budget.
They need to decide what to do first, how much to invest and how to measure it. I give them priorities and numbers, without the jargon. AI for executives and boards →
They ask about risk, data and AI governance. I prepare sessions on what the board should decide now and what is better left to wait. Sessions for boards of directors →
This is where I have the longest track record: more than 20 years in digital marketing. Agents that answer and qualify, sales funnel analysis, and plain-language queries over the CRM.
Carlos Betancur Gálvez provides artificial intelligence consulting from Medellín, Colombia, for companies in Colombia, Ecuador, Peru and Central America: Costa Rica, Guatemala, Honduras, El Salvador, Nicaragua and Panama. The work is done remotely (assessment, working sessions and follow-up over video call) and on site when the project calls for it, for example a workshop with the executive team or a session with the board.
He is based in Medellín and travels from there; visits are scheduled around the project calendar and travel. The whole region works in Spanish and on Colombian time or one hour apart, so meetings fit into a normal working day in any of these countries. Projects can also run in English.
As a speaker he has given talks in nine countries: Honduras, Colombia, Spain, Ecuador, Costa Rica, Aruba, Bonaire, Curaçao and Chile. If you are looking for a talk for an event, see the AI keynote speaker page or, for events in the region, the AI keynote speaker in Central America page →
A Colombian medical equipment company. We built it with the btodigital team and it is running right now.
There are more projects, presented by industry and under NDA, on the Claude implementation page.
It depends on what you need to solve. All three are useful, for different things.
One person who assesses, decides with you and answers for the judgment. Useful when you need clarity and priorities before building, or a second opinion on what you are being offered.
Large teams, methodology and corporate scope. A fit for multi-year transformations and for organizations that need that backing, with larger budgets and longer timelines.
A team that builds and maintains the solution: development, integrations and support. It is what you need when the project calls for hands as well as judgment.
In my case two of them come together: I handle the advisory work and the architecture myself, and when something has to be built, we implement it with the team at btodigital, my agency.
I don’t publish rates because every project is quoted by scope, after the assessment. The first 30-minute conversation is free, and it tells you whether your case makes sense for artificial intelligence and what order of magnitude the investment would be.
This page covers artificial intelligence consulting in general: what to do, with which tool and in what order, without starting from a model. The Claude consulting page is for companies that have already decided to implement Claude (Anthropic) and want to see the cases, the architecture and the process with that technology.
The ones running in production today: Claude, ChatGPT, Gemini and open models. The choice depends on the problem, the cost and how sensitive the data is, and many solutions combine several models. The recommendation comes out of the assessment, not out of a brand preference.
Yes. I am based in Medellín, Colombia, and the consulting is delivered remotely for companies in Ecuador, Peru, Costa Rica, Guatemala, Honduras, El Salvador, Nicaragua and Panama, in Spanish or English. When a project calls for it, I travel for workshops or on-site sessions, scheduled around the project calendar and travel.
All three. With executive teams and boards the work is about decisions: priorities, risk, investment and how to measure. With business and technology teams it is about design and execution. You don’t need an in-house technical team to get started.
The assessment takes one week. A proof of concept takes two to three weeks, and a production-ready implementation takes four to six. After that comes continuous improvement, monthly or quarterly, depending on what the company needs.
I handle the advisory work and the architecture myself. When something has to be built, we build it with the team at btodigital, my agency: developers, designers and data specialists, depending on the scope of the project.
Three paths depending on where you are. I answer personally.
A 30-minute session over WhatsApp or video call. I review your case and tell you whether it makes sense to keep talking.
Cases by industry, architecture and the process for implementing Claude (Anthropic) in a company.
Artificial intelligence keynotes and workshops for companies, industry associations and universities.