AI consulting for companies
Custom AI solutions connected to your real data, with people in charge of what matters. For companies that are past the testing stage and want results.
What AI can solve in a company
These are the capabilities I build to fit your business. Each one connects to your systems and is measured in hours saved or in sales.
Customer service
AI assistants that answer with your information and know when to hand the conversation over to a person.
Internal agents
Assistants that query your ERP, CRM and documents to answer your team with real data.
Process automation
Workflows that connect your tools and handle data transfers, notifications and reports on their own.
Document reading
Extract data from invoices, labels, contracts or images, and leave it organized in your system.
Analysis and alerts
Data brought together in one place and alerts before the problem hits your cash flow.
Marketing and content
Assisted production and publishing, with human review and results measured in sales.
Five rules to make AI actually work
Problem first
We pick the task that hurts, not the trendy tool. If the solution is simpler than AI, we start with that.
Real data
The AI queries your information, shows its source and admits when it doesn't know.
A person approves
Anything that affects customers, prices or money goes through someone on your team.
It gets measured
Every project starts with a number: hours, response time, errors or sales.
It stays yours
Your data and your rules stay in your company, with no vendor lock-in.
Buy, adapt or build?
Not everything needs custom development. Part of a good consulting engagement is telling you when it isn't necessary.
Buy a tool
Makes sense if your need is standard and you want to start quickly. You need to check how the cost grows with your volume.
Adapt what you already have
Makes sense if your systems work well and they just need to be connected or given AI at specific points.
Build custom
Makes sense if the process is part of how you compete. I covered this decision in Your own CRM versus SaaS and in SaaS isn't dying: it's dissolving.
Real cases, told from the inside
More reading in Drafts.
My AI-powered CRM served over 10,300 people. The secret: knowing when to stay quiet
All of an eCommerce store's channels on a single dashboard, with an AI assistant that replies and knows when to hand the conversation over to a person.
Read the case
My AI coworkers can't change anything without my permission. That's why I trust them
AI assistants that read real data from the ERP and the stores, cite their source, send alerts on their own and only act when a person approves the card.
Read the case
The product sold well and still expired in the warehouse. I built a system to see it in time
A system that cross-references each product's sales with each batch's expiration date and warns you in time which stock will expire unsold.
Read the case
I take a photo of the label and the system tells me which warning seals it carries and which ingredient isn't allowed
A single record per product: AI-read label, automatic seals and warnings, regulatory filings with history, and sync with the stores.
Read the caseFrequently asked questions
If we already pay for ChatGPT, why do we need consulting?
ChatGPT helps each person separately, but it doesn't know your data or work inside your systems. Consulting makes the AI query your real information, follow your rules and be part of the process, without relying on someone to copy and paste.
What's the difference between a chatbot and an AI agent?
A chatbot converses and answers. An agent also queries systems, prepares actions and carries them out, always within the permissions you define. What your company needs depends on the task: a good chatbot is often enough.
Where does a company that has never used AI start?
With a repetitive process that has a measurable cost and available data. Handling inquiries, sorting documents and weekly reports tend to be good candidates.
Is it better to buy a SaaS or build custom?
If your need is standard, a SaaS is usually the fastest route. Custom makes sense when the solution is part of how you compete, when usage costs grow along with your volume or when you need rules a SaaS doesn't allow.
How do you keep the AI from making up answers?
It's limited to defined sources, required to show where each piece of data came from and asked to admit when it doesn't know. For important actions, a person reviews and approves before they happen.
Do you work with companies outside Concepción?
Yes. I'm based in Gran Concepción, but consulting and development can be done online with companies across Chile.
Let's bring AI into your operation
Tell me which process you'd like to solve and I'll tell you where I'd start.
Let's talk about your project