Santiago Santa María Morales
Also known as: Santi Santa María · santismm · Santiago Santa María
Santiago Santa María Morales is Group Chief AI Officer at ACCIONA, where he leads the group's artificial-intelligence strategy. Previously he spent nearly a decade at Google (LATAM and the US), where he led the AI Forward Deployed Engineers team for Latin America, helping enterprises become AI-driven through Generative AI and Vertex AI. Earlier, during his years at NTT DATA, he led the Conversational and Agentic AI practice — scaling the Syntphony platform with a team of 50+ specialists across 7 countries and business on 5 continents — including pioneering work in social robotics with JIBO (MIT Media Lab) in Boston. A physicist (astrophysics and cosmology) trained at Universidad Autónoma de Madrid and Freie Universität Berlin, he has spoken at events in Madrid, Boston, Tokyo, Chile and across LATAM. He combines technical rigor with strategic vision to turn deep technology into business disruption and scientific progress.
Copy and paste it straight into your communications.
What I do, and where I write it
I'm Santiago Santa María Morales. A physicist by training (astrophysics and cosmology), today Group Chief AI Officer at ACCIONA. Before that I spent nearly a decade at Google (LATAM and the US) and, earlier, at NTT DATA, leading the Conversational and Agentic AI practice (Syntphony platform) across five continents.
EUREKA is where I write — frameworks, deep dives and conversations on AI for people who actually have to ship it. Everything I publish ends up in this library, no matter which channel it started on.
Beyond writing and speaking, I publish interactive tools and open benchmarks at SANTISMM Labs.
The frameworks, and where to read them
Each one has its own page, a JSON version where it applies, and a citation.
- Harness Engineering Handbook
The canonical long-form handbook on harness engineering — definition, taxonomy, memory, observability, evaluation, orchestration, security and governance — for humans and AI agents.
JSON - Agentic Control Matrix
18 agent controls mapped to the EU AI Act, ISO/IEC 42001, NIST AI RMF, OWASP LLM Top 10 and MITRE ATLAS — with a test for each. JSON and CSV.
JSON - The Harness Maturity Model
Seven levels from prompted to adaptive, each with criteria you can pass or fail — not descriptions you can agree with. Cumulative: a level is held only when everything below it passes.
JSON - The Harness Scorecard
One formula and eleven metrics for comparing agent harnesses, with the primary unit set at cost per correctly verified outcome rather than successful runs.
JSON - The Agentic Stack Map
One printable page: the 18 controls of the Agentic Control Matrix, grouped by where they act in the stack, with what each one answers to and how to test it.
- Taxonomy of autonomous agents
An interactive taxonomy that encodes 24 AI agents — from OpenAI, Anthropic, Google, Microsoft, Amazon and SpaceXAI to the open-source frontier — as vectors across six orthogonal axes plus an action surface, with a filterable table, an A×T×I governance-risk matrix and a machine-readable JSON API.
JSON
What I work on
Recurring areas across writing, talks and consulting.
Education & certifications
Where the physics and the cloud come from.
- BSc Physics — Astrophysics & CosmologyUniversidad Autónoma de Madrid1995–2003
- BSc Theoretical & Mathematical PhysicsFreie Universität Berlin1999–2000
- Cloud Digital LeaderGoogleCertification
- Leadership Academy — Modules I & IIGoogle2024
- Accelerate Hackathon — ParticipantGoogle2026
Where to find me
Speaking & events
Keynotes, panels and workshops connecting frontier technology — agentic AI, foundation models, conversational AI — with business decisions and scientific progress.