SCALE-Decode：解锁英国中小企业增长工作流博士研究员

Small and medium-sized enterprises (SMEs) underpin UK employment and innovation, yet many struggle to turn early success into sustainable growth. Rising demand can expose informal routines, disconnected systems, founder dependence and employee overload. Managers may see delayed delivery or lost sales without understanding how constraints travel across the firm. Research and policy treat issues separately, while generic AI lacks context. SCALE-Decode examines how responsible AI can reveal growth workflows, diagnose bottlenecks, and build an AI system to support managerial actions.

Work Package 1 (months 1-12) will map and explain SME scale-up workflows. Manager and employee interviews, walkthroughs, task diaries, process documents and privacy-preserving event logs, where suitable, will be integrated through qualitative analysis, Business Process Model and Notation and process mining. Cross-case comparison will distinguish symptoms, including delay and rework, from underlying causes such as centralised approvals, capability gaps and poor information transfer. Stakeholder validation will refine three deliverables: a Scale-up Workflow Framework, anonymised maps and a bottleneck taxonomy. Together, they will help firms and advisers trace local constraints across an organisation and strengthen workflow-centred scaling theory through credible evidence.

Work Package 2 (months 10-24) will build SCALE-Decode, a secure AI system to help SMEs scale by reconstructing workflows, diagnosing scaling bottlenecks and generating feasible action plans. A knowledge graph will connect activities, dependencies, organisational conditions, evidence and interventions. Four AI components will gather information, produce editable maps, diagnose constraints, and generate and critique actions. Each recommendation will identify its evidence, prerequisites, expected benefit, cost, effort, risk and uncertainty. Managers will correct maps and approve recommendations. Technical, adversarial, usability and unseen-case testing will assess fidelity, traceability, feasibility and unsupported claims. Deliverables comprise a validated prototype, documentation and responsible-use guidance.

Work Package 3 (months 22-36) will evaluate feasibility, usefulness and implementation in SMEs. A longitudinal field study will compare AI-assisted diagnosis with existing approaches and track workflow understanding, decision quality, action completion and selected operational indicators. Usage records, surveys and interviews will show why recommendations Engineering and Physical Sciences Research Council Doctoral Landscape Award are accepted, adapted or rejected and whether outcomes vary with workflow maturity, digital readiness or owner-manager dependence. Case studies will examine employee capability and wellbeing, recognising that gains must not depend on work intensification. Deliverables will include an evaluation dataset, comparative findings and implementation evidence, establishing where SCALE-Decode adds value, where human expertise remains essential and which conditions support responsible adoption.

Work Package 4 (months 36-42) will integrate findings and translate them for research and practice. Outputs will include academic papers, reusable workflow passports, sector templates, checklists, an adviser toolkit, policy brief and public-engagement materials. Practitioner workshops will refine resources and inform future development. This package will convert evidence into usable guidance for SMEs, advisers, growth hubs, policymakers and technology providers.

Funded by: EPSRC Engineering and Physical Sciences Research Council £21,805. The studentship will provide funding of fees and a stipend of £21,805 per annum for 2026/7.