Jose SantosJOSE SANTOS' APPLIED AI
How I Apply AI, from Business Problem to Working System
I use AI where it can improve a consequential product, decision, or operating workflow. My work begins with the outcome and the people accountable for it, then connects product strategy, process design, data, software, automation, and human control to produce something that can be used and trusted.
My AI Work at a Glance
Enterprise product leadership
Routing and scheduling data ingestion
Hands-on product development
Photo World Clock transformation and launch
Release and operating automation
Reusable App Store release workflows
Governed AI workflows
Ownership, validation, human approval and traceability
Turning recurring customer data mapping into a reusable AI workflow
Case 1 / Routing and Scheduling Solution
Director of Product Management at an enterprise software company
Observed outcomeRecurring data-preparation delays emerged in work with numerous prospects and two paying customers. In one directly observed paying-customer implementation, after the capability was released, the data-preparation stage, which represented most of the onboarding effort, was reduced to near-zero ongoing manual setup once the reusable mapping protocol was established.
The problem
Repeated field mapping and transformations across varied customer sources.
My role
Product strategy, opportunity, requirements, accuracy boundaries and cross-functional delivery.
The capability
Human-readable mapping instructions become reusable protocols connected through MCP.
Product decisions that mattered
- Start with the onboarding constraint and define the required accuracy before selecting the capability.
- Turn human-readable instructions into a precise, reusable mapping protocol.
- Support recurring imports and changes in customer systems or formats.
- Keep accountable review around consequential transformations.
- Measure the change in customer setup effort.
Related platform delivery context
The broader routing and scheduling platform moved from business plan to General Availability in under nine months and acquired three customers in its first quarter. These are platform delivery and commercial results; they are not attributed solely to the AI ingestion capability.
Rebuilding and launching an iOS product with an AI development partner
Case 2 / Photo World Clock
Independent product owner leading strategy, design and software delivery through release and go-to-market
OutcomeReleased August 2026 after a four-week transformation and launch cycle. The product became free to download with optional in-app purchase, supported by reusable localized release workflows.
Product and business decision
Repositioning, view rearchitecture and a change from paid-only to free with optional in-app purchase.
How I worked with AI
I used Claude Cowork as a development partner while owning product direction, design and implementation decisions, testing, release quality and launch.
Localization workflow
I built and used a private Localization GPT with product-specific instructions, a terminology glossary and structured spreadsheet inputs to preserve context and consistency.
Scope of the transformation
- Repositioned the product as Photo World Clock and reworked navigation between photos and locations.
- Adopted the iOS 26 Liquid Glass interface and updated playback, Smart Fill and widgets.
- Implemented free download with optional in-app purchase, product analytics and release preparation.
- Prepared localized copy, screenshots, metadata and the go-to-market package.
Release automation and controls
The documented tooling produced 540 upload-ready App Store screenshots and 360 localized website images. It covers three device classes and 39 capture locales, including 36 App Store localizations and three website-only locales.
Validation checks dimensions, color mode, transparency and checksums. Dry-run previews and scoped replacement precede external uploads; unchanged items are skipped and interrupted transfers can recover. App Store metadata tooling maintains five content areas across 39 locale variants.
Operating Problem and AI Pilot Canvas
A working framework for assessing an AI workflow. Use it to clarify the operating outcome, decide what belongs with people and systems, and define the evidence needed before a pilot scales.
Understand the problem → Choose the approach → Design the test → Make the decision
Decision summary
Choose one workflow. Define a bounded pilot. Agree on evidence.
Product Strategist GPT
Your Virtual Product Manager
Product Strategist GPT brings my approach to product management into a tool product managers can use in their daily work. I made it available to help clarify decisions, challenge assumptions, and move from strategy and positioning to roadmaps, requirements and go-to-market plans. It works through tradeoffs using customer evidence, commercial reality and delivery constraints, helping you turn a question into a clear next step.
I use it to assist my work. Give it a try.
Assign it a task from your own work:
- Challenge my roadmap against the business objective, customer evidence, and delivery constraints.
- Compare these positioning options and recommend one, including the commercial and go-to-market tradeoffs.
- Turn this product concept into a commercially grounded PRD with explicit exclusions and acceptance criteria.
Active work: building new software businesses
I am actively developing two software ventures, using Codex and AI-assisted workflows to connect product definition, implementation and review. Each has a different scope and delivery model.
Delivery operations: a structured path to a pilot
I am doing most of the work directly, organizing delivery into bounded stages with requirements, implementation tasks and acceptance checks. I use Codex to develop and refine the software and its operating documentation. Driver, dispatcher and platform-owner guides help expose workflow gaps, while scenario-based validation examines conditions such as lost connectivity. The preliminary driver mobile apps and web interface for data intake were produced in three days, from specifications to code. I am now validating the workflows and closing gaps toward a pilot release.
Risk management: coordinating a complete platform
I am working with a founding team toward a supply chain risk-management platform. I am using spec-driven development (SDD) to organize AI-led development around explicit requirements and acceptance criteria. I structure the work so a product manager, a supply-chain subject-matter expert and software engineers can direct the creation of the platform using AI. AI-assisted work supports the preparation and reconciliation of requirements, evaluation criteria, data definitions and review materials. I structure the effort around shared user journeys, named responsibilities, versioned working sets and collective review gates so that product, domain and engineering decisions stay aligned. Current work includes platform definition and review; a complete platform launch is the objective.
Put my experience to work
If you are hiring a product leader or looking for experienced support on a defined initiative, let’s discuss what you need to move forward.