From Public Data to MXN 18.45 Million in SME Credit: An End-to-End Commercial Pipeline

Context

Between October 2025 and March 2026 I was an SME Key Account Manager at Konfío, placing credit for small and medium-sized businesses. My bet was not to depend on isolated contacts, but to build a system that generated opportunities in an orderly, measurable way.

The territory was all of Mexico, with a focus on SME profiles. By agreement with management, I concentrated on the northwest of the country (Sinaloa, Sonora and the Baja Californias) and on the supply chains with the best credit performance at Konfío, among them construction and its specialty trades.

The challenge

An SME does not show up on its own: it has to be found, its profile understood, contacted cold, and accompanied through to signing. Doing that with manual tools limits both volume and the quality of follow-up. My challenge was to turn that process into a structured, repeatable flow built on data of my own.

My role

I designed and ran the whole pipeline. It was mine from start to finish: from identifying the data source to the disbursement of the credit.

What I did

I built a data-capture and transformation pipeline and connected it to a complete sales process:

The pipeline had two entry points into the same route: some companies came in directly and registered through self-service, and others did so after a cold call. The steps were:

1. Integration of public-source data: I designed a Python script with AI, tested and run in Google Colab. I integrated Ollama and AI models to collect data from more complex sources.
2. Cleaning and transformation: the ingestion system produced a structured Excel database, usable as a database of my own.
3. Segmentation with business criteria: an agent split the database into small batches to be worked under anti-spam rules. It segmented them by Konfío’s indicators, that is, the sectors that approved the most credits, and by their fit with supply chains such as construction and its specialty trades.
4. Own campaigns: I applied copywriting techniques to encourage self-service, using a structured link that carried my details as Key Account Manager.
5. Measurement and refocusing: the gauge was a response within 36 hours, either a reply or a registration. When a sector responded, I refocused effort on it and continued with the sequence of emails and cold calls.
6. Registration and evaluation: companies registered themselves in Konfío’s database, integrated with Salesforce, or did so after a cold call, and went through the credit evaluation process.
7. Closing, signing and disbursement: I accompanied each deal through to contract signing and credit disbursement.

Result

MXN 18.45 million in SME credit placement between October 2025 and March 2026, the result of a pipeline of my own that joined data and sales end to end.

What I still apply

A sales process is also a data problem: if the source, the cleaning and the follow-up are well designed, the result stops depending on luck. That is the logic with which I connect operations, business and technology today.

Specific cases within this experience

A database of my own, built from public sources

• Situation: prospecting depended on finding companies one by one.
• Decision or action: I designed a Python script with AI, tested in Google Colab, that integrates Ollama and AI models to collect data from complex sources, and I turned the data into a structured Excel database.
• Result: a database of my own, segmented by the sectors with the most approvals at Konfío and by supply chains, which made it possible to measure which campaigns worked.
• To discuss: how to turn scattered public information into a usable commercial database, and how to segment it responsibly.

From cold contact to disbursement

• Situation: each company travels a long road from first contact to disbursement, with many points where it can be lost.
• Decision or action: I used a response within 36 hours as the gauge for each campaign: when a sector responded, I refocused effort on it and continued with emails and cold calls. Some companies registered through self-service and others after a call, and I accompanied all of them through to signing and disbursement.
• Result: MXN 18.45 million in placement.
• To discuss: how to design a B2B sales process where the company moves forward on its own and every stage can be measured and improved.

Let’s talk about

If your organization wants to join data and sales, or to build a database of its own to sustain a B2B sales process, I would be glad to share how I solved it at Konfío.