CloseAlly

AI consulting + product engineering

AI advice.
Working
software.

For product owners and business leaders at established U.S. companies. Advice, workshops, and hands-on engineering, from the first question through ongoing support.

Talk through your project
Led by Luis HernándezAI platforms, privacy copilots, and operational assistants.
Engineering lead · Guadalajara, MexicoExplore the engineering experience
Illustrative workflowSample data
Review email changes before updating a customer recordIllustrative workflow with sample data. An email asks Customer Success to take ownership of the handoff and follow up Tuesday. AI matches Sample Account and prepares source-linked changes. A person reviews the proposal before saving. A second email proposes Friday. A person reviews that change before it is saved. A customer conversation, connected.Email → your CRM Incoming emailAccount handoff“Customer Successwill own the handoff.Follow up next Tuesday.”Source: original email AI preparesan update Your customer recordSample AccountHandoff ownerCustomer SuccessNext follow-upTuesday → FridayEmail evidence attached A person reviewsApproved, then saved Then, a second email“Make it Friday.” Keep the source email with the saved update.Email becomes a reviewed customer recordIllustrative workflow. AI matches an email to Sample Account and keeps the source. A person approves the proposed changes before saving. A second email moves Tuesday's follow-up to Friday, with another human review. Account handoff“Customer Success owns it.Follow up next Tuesday.”Source: original email AI prepares an update A person reviewsApproved, then saved Your customer recordSample AccountHandoff ownerCustomer SuccessNext follow-upTuesday → FridayEmail evidence attached Next email: “Make it Friday.”
Review email changes before updating the customer record.

Selected experience

AI engineering,
inside the product.

Work from Luis's engineering practice, including his own platform and contributions within broader product teams. Each example focuses on what he built.

Privacy automation copilot

Connect the model
to the whole application.

The copilot needed custom retrieval and conversation processing to work across the interface, backend services, and model integration. Luis led full-stack development alongside technical leadership, product, and design, including a custom chat-processing protocol.

Luis's contribution

Hands-on technical leadership and implementation. Delivered a usable application with code review, tests, and deployment checks.

AI assistance for laboratory software

A conversation needs
a working session.

The assistant needed to keep a conversation connected to its execution session and recover that session when work resumed.

Luis built core platform features connecting persistent conversations, isolated execution, session recovery, and browser integration.

Luis's contribution

Core platform engineering within a broader product team. Delivered connected conversation and execution features, with authentication and end-to-end tests around recovery and user isolation.

Internal engineering · Personal platform

An AI platform that can pick the work back up.

AI tasks often reach beyond a single conversation. Luis built a personal agent platform connecting private tools with scheduled and resumable work.

Saved state and human review are part of the execution flow, so a task can pause for a decision and continue with its context. Checks covered restart recovery and task continuation.

Luis's contribution

Platform architecture, tool integration, and workflow implementation.

How we help

Choose the next step
your team needs.

Some teams need help deciding where to begin. Others have a prototype to finish or an AI-heavy application to maintain. The engagement starts with your situation.

AI consulting

Find a practical direction for AI.

Work through opportunities, tools, and constraints with someone who can also implement the result. Turn a broad adoption question into an approach your team can evaluate.

Bring a business task, a product question, or an idea that needs a second look.

Team workshops

Help your people use the tools.

Practical learning for engineers and nontechnical teams, built around the work they actually do. Explore AI-assisted development, everyday workflows, and how to review the output.

Start with real examples from the team and a clear learning goal.

Build or improve a product

Get from a promising idea to working software.

Build a new application, add AI to an existing product, or improve the connections between systems. A focused first scope gives you something concrete to review.

That could be a prototype, an integration, or a defined product change.

Maintain AI software

Keep the product understandable and usable.

Maintain existing applications, including AI-heavy products and software built with coding agents. Review the code, resolve issues, and plan improvements around how the product is used.

Support and maintenance plans are scoped to your product and operating needs.

A support concept

Support can start in the tools people already use.

A person asks for help in their AI assistant. A connected tool prepares the request for your ticketing system. Your team handles triage, and the status comes back to the conversation.

The connection behind it

MCP can expose a support tool to ChatGPT or Claude. The implementation would define permissions, the information a ticket needs, and how status is retrieved. This illustrates a proposed integration.

Support conceptIllustrative workflow
Support starts inside the conversationIllustrative support workflow. A request in ChatGPT or Claude passes through an MCP support tool into ticketing. A human triages it. The status returns to the original conversation. ChatGPT / Claude“Can you help withthis order?” MCP connectionYour support tool Support ticketRequest + contextReady for triage Your teamHuman triage Status back in the chat Return the ticket status to the original conversation.Support from the conversation to the teamIllustrative support workflow. Someone requests help in ChatGPT or Claude. A support tool connected through MCP prepares a ticket. A human triages it and its status returns to the conversation. ChatGPT / Claude“Can you help withthis order?” MCP connectionYour support tool Support ticketRequest + contextReady for triage Your team triagesA person takes it on Status back in the chat
Send a support request from the chat and return its status.

How we work

A small team,
close to the decisions.

A focused team of one to three developers can work with your product owner or project manager. CloseAlly can also provide product leadership when the engagement needs it.

Discuss how we'd work together

Start with a defined question

Clarify the business task, existing systems, and the uncertainty the first scope should address. Agree on responsibilities and the result to review.

Keep the conversation open

Weekly or biweekly updates and a shared channel keep product questions connected to the implementation.

Plan for the software's next stage

Discuss documentation, handover, and maintenance as part of the scope. Support and further development can continue under an agreed plan.

Engineering beyond the application

Deployment and infrastructure work can include AWS, Google Cloud, Vercel, and managed serverless services, with infrastructure as code and delivery automation where they fit.

Luis Hernández, engineering lead

Your engineering lead

Luis Hernández

Hands-on engineering.
Product decisions included.

Luis is a software engineer and technical lead whose experience spans AI assistants, operational platforms, and full-stack products.

His work connects architecture and implementation with the practical details of delivery: clarifying requirements, guiding developers, integrating systems, and resolving production issues.

Talk with CloseAlly

Before we talk

A few practical
questions.

We’re considering AI. Where should we start?

Start with a task your team understands and examples of how it happens today. Advice or a workshop can help identify where AI is useful, what needs testing, and whether there is a worthwhile software change to build.

Can you work with our existing team?

Yes. We can integrate with your product owner or project manager and work alongside your engineers. The proposal sets out responsibilities, the review process, and how the collaboration will run.

Our application was built with AI tools. Can you maintain it?

We can review the application and the systems it depends on, then discuss a maintenance or improvement scope. Access, the state of the code, and operating requirements shape the plan.

How do you evaluate an AI feature?

Define the behavior the task needs and test representative examples. The review should include successful behavior, failure paths, and the decisions that need a person.

What will it cost, and how long will it take?

That depends on the task, existing systems, and the result you need to evaluate. We discuss scope, timing, and commercial terms in a proposal before work begins.

Can our team take over the software?

We discuss who will operate and maintain the software, then include the relevant documentation and handover requirements in the scope.

Let's get specific

What would help your team
move forward with AI?

Start a conversation