− POLICY Generative AI (GAI) promises superior analytics and agility in strategy work, yet organisations struggle to m…± BUSINESS The agricultural sector, a vital industry for human survival and a primary source of food and raw materials,…+ HEALTH With the advent of Large Language Models (LLMs), medical artificial intelligence (AI) has experienced substan…+ BUSINESS Recently, Gen AI has garnered significant attention across various sectors of society, particularly capturing…− CRIME Abstract Artificial intelligence (AI) has become integral to financial institutions operations. Implementing…± POLICY Artificial Intelligence (AI) has reshaped consumer contracts by enabling automated drafting, personalization,…− LABOR Abstract This paper examines the relationship between artificial intelligence, contemporary artistic producti…+ POLICY In February 2024, the Council and the European Parliament (EP) agreed on the Artificial Intelligence Regulati…
TruaceTracing the truth around AIMonday, July 20, 2026
Subscribe
Policy·P Space·Evidence-backed problem·Published 2026-07-20

From pilots to decision systems: embedding generative AI into strategic decision-making through a socio-technical and governance lens

Generative AI (GAI) promises superior analytics and agility in strategy work, yet organisations struggle to move beyond pilots towards routinised decision inputs. This study investigates how GAI becomes embedded in strategic decision-making (SDM) through a qualitative single-case analysis of a global multi-brand group, based on 27 semi-structured executive interviews triangulated with internal documents and industry reports. Structured inductive coding yields a process model identifying enablers, leadership-driv…

TRV-2026-0400Peer-reviewedPermanent record — cite & verify
From pilots to decision systems: embedding generative AI into strategic decision-making through a socio-technical and governance lens

"25.5.2012 Seminario 'Medidas sociales y políticas sobre la familia' / 5.25.2012 Panel Session 'Family Social and Government Policies'" by HazteOir.org is licensed under CC BY-SA 2.0. To view a copy of this license, visit https://creativecommons.org/licenses/by-sa/2.0/.

The quick read

Researchers conducted a qualitative single-case study of a global multi-brand group to examine how generative AI moves from pilots to routinised inputs in strategic decision-making. Based on 27 executive interviews and document triangulation, they identified enablers and barriers and proposed a four-stage pathway from Awareness and Exploration to Institutionalisation and Transformation with quality, provenance, explainability and accountability gates.

The work matters because it reframes generative AI from an autonomous oracle to governed decision support and offers heuristics for scaling responsibly. Uncertainty remains about transferability beyond the single case, long-term performance of governance gates, and whether quick wins translate into sustained strategic transformation.

Main points
  • Qualitative single-case analysis based on 27 semi-structured executive interviews triangulated with internal documents and industry reports.
  • Identified enablers: leadership-driven adoption, quick wins, prompt-based experimentation, workforce training, secure platforms, dedicated investments.
  • Identified barriers: strategic ambiguity, limited awareness, hallucination risks, prompt-engineering deficiencies, data readiness, privacy or IP concerns.
  • Proposed four-stage pathway: Awareness and Exploration, Experimentation and Pilots, Formal Adoption and Integration, Institutionalisation and Transformation.
  • Governance mechanism includes admission gates for quality, provenance, explainability, and accountability with human-in-the-loop redistribution of responsibilities.
Problem

Organizations attempting to embed generative AI into strategic decision-making struggle to move beyond pilots due to hallucination risks, prompt-engineering deficiencies, data readiness gaps, and privacy or IP concerns.

The rundown

The study used structured inductive coding of 27 executive interviews plus internal documents and industry reports to build a process model of GAI embedding.

The model specifies four stages with admission gates for quality, provenance, explainability, and accountability, and describes human-in-the-loop arrangements that redistribute responsibilities between AI and managers.

Governance templates and socio-technical alignment were found to determine whether GAI outputs are admitted into formal deliberation, reframing GAI as decision support rather than autonomous oracle.

Sources

Reader signal

How should this claim be treated?

The debate