All work

In production

AI-Assisted Translation & Curriculum-Adaptation Workflows

~4 months ~2 weeks

Translation turnaround at Santillana, with costs ~90% lower.

Streamlined AI-assisted translation workflows for a publisher (turnaround ~4 months to ~2 weeks, costs ~90% lower) and supported the automation of curriculum adaptation (~50% faster, ~95% cheaper, delivered in-house).

Built at
Santillana (Sanoma Group)
For
Santillana (Sanoma Group)
Period
May 2024-Jul 2026

Context

Santillana is part of Sanoma Group. From May 2024 to July 2026 I worked there as an AI & Automation Consultant.

Architecture

  • I prepared editorial content for AI-assisted processing and retrieval: a new AI-friendly dataset built from proprietary original editorial content and previous content, to reduce friction with AI tools. Techniques named: RAG, chunking, LLM-based tagging, knowledge graphs and data structuring.
  • I streamlined AI-assisted translation workflows.
  • I supported the automation of curriculum adaptation, bringing delivery in-house.

Components in the diagram. Hover or focus one to see its connections.

Rollout

  1. Status: the workflows are still in use at Santillana, after my role ended in July 2026.

  2. I advised and trained editorial teams on adopting generative AI and workflow automation. The training also covered AI agents and emerging AI technologies, supporting their adoption across business and editorial workflows.

Impact

  • Translation turnaround: from ~4 months to ~2 weeks.
  • Translation costs: ~90% lower, with potential savings of up to 98 to 99% vs. historical budgets.
  • Curriculum adaptation: processing time ~50% lower and costs ~95% lower, with delivery brought in-house.
  1. Translation turnaround~4 months ~2 weeks
  2. Translation costs~90% lower
  3. Potential savings vs historical translation budgetsup to 98 to 99%
  4. Curriculum-adaptation processing time~50% lower
  5. Curriculum-adaptation costs~95% lower

Discovery

Stakeholders