Sofpact field notes

AI strategy and operations, without the theatre.

Analysis for leaders working through use-case choice, governance, operating models, data readiness and adoption.

How these notes are made

Written with AI assistance. Facts link to the primary sources named in each note, and Radko Diev decides what is published.

Browse all published field notes

Editorial documentation chain from source evidence to approved record beside a continuous audit trail
13 September 2026 · 7 min read

AI in Regulated Life Sciences: Every Claim Needs a Traceable Record

AI in regulated life-sciences documentation: defensible attribution, retained records, pinned model versions and a narrow intended use.

Read field note
Editorial timeline showing high-risk and product deadlines moving to 2027 and 2028 with obligations unchanged
11 September 2026 · 8 min read

EU AI Act High-Risk Delay: What Changed, What Did Not

High-risk dates moved to 2 December 2027 and 2 August 2028. The obligations did not: the inventory, roles and evidence to prepare in the meantime.

Read field note
Editorial three-step adoption staircase from one workflow to evidence to an operating model under EU rules
8 September 2026 · 7 min read

AI in the CEE Mid-Market: Sequence Adoption to the Team You Have

EU AI obligations apply at every size. Use regulatory sandboxes and SME measures, then sequence one workflow, evidence in context and an operating model.

Read field note
Editorial timeline showing three AI Act duties in force from 2025 and 2026 and two ahead in 2027 and 2028, with the system inventory as the starting point
5 September 2026 · 8 min read

The EU AI Act in Bulgaria: What Binds Your Bulgarian Entity Now

Which EU AI Act obligations already bind a Bulgarian entity, what arrives in 2027 and 2028, when a deployer becomes a provider, and where preparation starts.

Read field note
Editorial diagram showing value, controls and ownership converging into an AI strategy
3 June 2025 · 7 min read

How to Build an AI Strategy That Survives Contact with Operations

Build an AI strategy around valuable workflows, proportionate controls, decision-grade evidence, operating ownership and clear stop conditions.

Read field note
Editorial network map showing an AI agent connecting identity, permissions, payments and regulated services
7 April 2025 · 8 min read

AI Agents and Delegated Authority: The New Control Point in Regulated Services

A control framework for acting AI agents covering identity, delegated authority, payment permissions, policy gates, evidence, oversight and recovery.

Read field note
Editorial document-processing flow from source file through extraction, validation and action
29 January 2025 · 7 min read

Document AI or Generative AI? Choose the Architecture, Not the Hype

Choose a document AI architecture across OCR, extraction and generative interpretation, with evaluation, provenance, review and data protection.

Read field note
Editorial readiness grid for data purpose, permission, quality, provenance, evaluation and ownership
9 January 2025 · 8 min read

AI Data Readiness: A Practical Checklist for Managers

Assess AI data readiness across purpose, permission, quality, provenance, representation, evaluation, operating limits and accountable ownership.

Read field note
Editorial architecture for an internal AI assistant grounded in approved company knowledge
21 December 2024 · 7 min read

Internal AI Assistants: Ground the Knowledge Before You Scale

Build a governed internal AI assistant around approved knowledge, inherited access, representative evaluation, visible evidence and accountable ownership.

Read field note
Editorial retail floor plan with zones, flows and decision gates
14 December 2024 · 7 min read

AI for Retail Decisions: When Simulation Is Evidence

Use AI, computer vision and simulation to test retail layout decisions while separating assumptions, model output, evidence and accountable judgement.

Read field note