Beyond plug-and-play: How to adopt AI without losing expertise

Original paper: AI adoption as technology transfer: an integrative review and conceptual framework for governance, validation and knowledge preservation

Authors: Musarat Kabir-Chisty

Read the paper: https://doi.org/10.1007/s43681-026-01296-9

About the researcher

Musarat Kabir, Managing Director, Cloutech Innovations, UK. Find out more here: https://www.linkedin.com/in/musarat-kabir-chisty-phd-cmgr-mcmi-514a8155/

What problem does this paper address, and why does it matter?

The paper addresses a hidden risk in AI adoption: organisations may automate a workflow without fully understanding or preserving the human expertise that makes it work. Written procedures rarely capture every judgement call, exception or contextual cue that experienced professionals rely on. These can be lost when work moves into an AI-enabled system.

This matters because a system can appear efficient while overlooking knowledge essential to reliable decisions. Governance and human oversight become less effective if the underlying process was transferred incorrectly. The paper proposes treating AI adoption as a disciplined technology-transfer process, requiring evidence that expertise, performance and accountability are preserved before deployment and maintained over time.

What did this research discover/create?

This research developed the AI Technology Transfer Framework (AITTF), a seven-stage approach to helping organisations move human expertise and working processes into AI-enabled systems without losing essential knowledge or accountability.

It brings together research on organisational knowledge and AI governance with established practices from regulated industries such as pharmaceuticals and biotechnology. The framework covers understanding the work, capturing expertise, transferring processes, validating results, testing consistency, maintaining human oversight and continuously checking performance.

Two anonymised examples from legal and corporate services illustrate its application. The central argument is that successful AI adoption depends on verifying that the knowledge behind the work has been transferred correctly.

How could this research impact real-world applications?

The framework could help organisations turn AI pilots into dependable operational systems by making knowledge capture, testing and accountability part of implementation from the outset.

For example, an immigration practice used the approach to organise document handling while retaining professional review and escalation. A corporate-services practice consolidated fragmented company records into a traceable register checked against authoritative external information.

More broadly, the approach could help organisations identify missing expertise, define acceptable performance, clarify when AI must defer to people and detect problems after deployment. These case studies illustrate practical applications; they do not establish independently verified or universally applicable performance gains.

Who should care about this work? 

Business leaders, AI implementation teams, domain experts, risk and compliance professionals, policymakers and researchers should care about this work.

It is particularly relevant to organisations introducing AI into professional or regulated services, where decisions depend on experience, contextual judgement and exception handling. It also matters to the people overseeing these systems and to clients whose outcomes depend on their reliability.

What is noteworthy about this research?

The research draws attention to a question that can be overlooked when adopting AI: have we accurately captured the expertise behind the task we want the system to perform?

Experienced professionals often rely on knowledge that is never fully written down, including how to recognise exceptions, interpret context and judge when something needs closer attention. That knowledge can be lost during implementation.

The distinctive contribution is bringing established technology-transfer disciplines into AI governance, making the preservation and faithful transfer of expertise something organisations must demonstrate with evidence.

What's the ONE key takeaway you want people to remember? 

When organisations adopt AI, they are transferring expertise as well as tasks. They need evidence that the knowledge, judgement and safeguards behind the work survive that transfer.

Read More

Access the full paper: https://doi.org/10.1007/s43681-026-01296-9

Have research to share or recommend? https://forms.gle/kaTJEd3gfJV8vtDx8