AI Automation Services for Small Businesses
Data Quimbaya is an automation agency for small businesses. I build the workflow automation that runs the parts of your operation nobody enjoys doing: AI agents that answer customers on WhatsApp, sales follow-up that goes out on its own, CRM integration so records stop being retyped, and the API integration that makes tools which were never designed to talk to each other behave like one system. Built on n8n, with Zapier or Make where they genuinely fit, and custom code where they do not.
Embedded, not handed off
Most automation projects fail the same way: someone writes a 40-page process document, builds against it, and discovers on launch day that the document described how the work is supposed to happen, not how it actually happens.
So I do the opposite. Before anything gets built, I sit with the people doing the job and watch them work — the real WhatsApp thread, the real spreadsheet, the workaround nobody wrote down. That practice is called shadowing, and the engineering model built around it is known as forward deployed: the engineer works inside the customer's operation rather than receiving a spec from it. Palantir invented the role, OpenAI kept the name for the engineers it embeds with the companies deploying its models, and Microsoft runs the same idea under solutions engineering.
It is a model built for enterprise budgets. Running it as one person, for a shop with four employees, is the whole point of this business.
I watch before I build
A week of looking at how the work actually flows. Almost every time, the expensive problem turns out not to be the one you called me about.
You get a system, not a deliverable
No PDF of recommendations for you to implement. What ships is running software, connected to your real tools, with your real data moving through it.
One person, start to finish
No account manager relaying messages to a developer you never meet. The person who watched your operation is the person who writes the code and the person who answers when it breaks.
It keeps working after launch
Automations break quietly when a tool changes its rules. Monitoring and fault fixing are part of the monthly plans, not an incident you find out about from a customer.
Process automation
I look at how a key process runs today (quotes, welcoming new clients, reports, payment reconciliation) and build the flow that does it by itself, with the validations and checkpoints it needs to run without anyone pushing it every day.
Businesses repeating the same manual work several times a week that want those hours back.
- New-client welcome with documents that generate themselves
- Payment reconciliation across your payment provider, bank, and accounting
- A weekly report that builds itself and arrives on Monday
- Your CRM, billing, and support always in sync
n8n · Make · Zapier · custom code
Application integration
I make your CRM, your billing, your store, and your spreadsheets talk to each other. When a ready-made connection doesn't exist, I build the one your business needs.
Teams with 5 or more tools where someone is still copy-pasting between tabs.
- HubSpot, Notion, and Slack working as one
- Your store connected to inventory in real time
- Every booked call creates the contact and notifies the team
- Older systems connected to modern tools
n8n · Zapier · custom connections
AI lead qualification
Every prospect that comes in gets researched and evaluated with AI against what matters to you: size, urgency, and fit with your business. It lands in your inbox with a summary and a recommendation. You decide who to contact first.
Sales teams getting more prospects than they can serve well, needing to prioritize without losing quality.
- Automatic scoring with clear reasons
- A first-message draft with each prospect's context
- Every prospect routed to the right salesperson
- Instant alerts when a high-priority prospect arrives
Claude · GPT · HubSpot · Pipedrive
AI knowledge assistant
I build an assistant trained on your documentation (Notion, Drive, PDFs, contracts). It answers team questions, helps customers, and handles first-level support. It always cites the source so you can trust the answer.
Teams losing hours answering the same questions, or businesses wanting support that knows their product deeply.
- Internal support: "how do I approve an expense?", "what does my contract say?"
- A support chat that knows your product and documentation
- An internal-policy assistant for HR or legal
- Find any detail in old contracts, proposals, or reports
Claude · OpenAI · n8n
Automated reports and dashboards
I connect your sources (CRM, billing, sales, ads), bring the data into one place, and deliver dashboards and reports that arrive in your inbox on their own. No more hand-building the Monday spreadsheet.
Owners and leadership making decisions from data split across six different places.
- A dashboard with your business numbers in real time
- Automatic financial reporting: income, spend, cash
- Alerts when something moves outside the normal range
- Operations tracking: response times and open items
Looker Studio · Metabase · Postgres · n8n
WhatsApp agent
I design and ship agents on WhatsApp Business or web chat that understand context, reply with your business information, book appointments, and hand the conversation to your team when a human is needed. Every conversation is measured.
Businesses that serve customers on WhatsApp and lose sales by replying late, or that want to sell around the clock.
- A quote builder that drafts a basic proposal in the chat
- 24/7 support that hands complex cases to a human
- Automatic reminders and confirmations for appointments, payments, and deliveries
- Ads that lead to WhatsApp with instant qualification
WhatsApp Business · AI · n8n
Document automation
I connect your CRM or database to smart templates so every document (quote, contract, proposal, invoice, monthly report) generates itself with the right information, signature included, and reaches the client without anyone copy-pasting 30 fields.
Businesses where someone still opens a Word doc, copies data over, and edits 30 fields for each new quote or contract.
- Sales proposal auto-generated from call notes
- Contract and terms generated when a deal closes in the CRM
- Recurring invoicing with per-client personalization
- Monthly client reports tailored per account
DocuSign · Google Docs API · Templated · PDFKit · n8n · Notion
Pick the piece that hurts most
Each one works on its own and every one of them feeds the next.
AI WhatsApp agents
A conversational AI agent on the official WhatsApp Business API: answers instantly, qualifies, books into your calendar, and escalates to you with the full context.
See the WhatsApp agentQuoted per projectSales and lead follow-up automation
Automated follow-up sequences, lead scoring, CRM integration, and reminders, so a lead that went quiet gets chased without anyone remembering to do it.
See the follow-up systemFrom $797Business websites
A fast, indexable site built to get found on Google and messaged on WhatsApp. Not a template with your logo dropped into it.
See how websites work$97The Automation Map
A fixed-price business process audit: the three processes worth automating first, with costs and priorities. Credited in full toward the project.
See the MapWhat people ask before hiring an automation agency
What exactly does an automation agency do?
It takes the repetitive work your team does by hand — copying data between apps, answering the same questions, chasing leads, assembling reports — and turns it into workflow automation that runs on its own. In my case that means AI agents on WhatsApp, automated sales follow-up, CRM integration, and the API integration that connects tools which do not talk to each other. The work is called business process automation; what you notice is that nobody is retyping anything anymore.
Do you use n8n, Zapier, or Make?
n8n by default: it is self-hosted, so your data stays on infrastructure you control and the cost does not climb with volume. Zapier and Make are the right call for a small number of simple triggers where the monthly fee is cheaper than the setup, and I will say so when that is the case. Where none of the three fit, I write the integration directly against the API.
How small is too small?
If you have at least one process that eats a few hours a week and follows rules you can explain out loud, it is worth automating. Most of my clients are between two and twenty people: salons, workshops, contractors, immigration agencies, small clinics. Big enough that the repetition costs real money, small enough that nobody has time to fix it.
What does it cost and how long does it take?
Every price is published: the Automation Map audit is $97, a website starts at $797, a complete WhatsApp AI agent starts at $997, and full implementations that connect several pieces land between $1,000 and $5,000. Monthly plans that keep everything hosted and monitored run $150 to $600. Most projects ship in two to four weeks. Your exact number is fixed and in writing before anything gets built.
Do you work in English or Spanish?
Both, natively, and that matters more than it sounds. If your customers write to you in Spanish and your suppliers invoice you in English, the agent has to handle both without switching personality. I serve clients across the United States on US hours, from a US WhatsApp line.
Who actually does the work?
I do. Data Quimbaya is DataQuimbaya LLC, and the founder is Lizandro Ruiz. There is no team of subcontractors between you and the person writing the code — that is the whole reason the embedded model above is possible at this price.
Every week you wait leaves money on the table.
In 30 minutes I'll show you what to automate first, what it costs, and when it pays back. Free.