AI Delivery Track Record:
10 to 15 years across data, AI and engineering, including 4 or more years leading technical teams. A record of taking AI, ML or GenAI solutions into production at enterprise scale, with specific examples of what went live and the benefit delivered.
Hands-On Technical Contribution:
Still builds. Able to prototype, write code and work a technical problem through personally, and to review design and code credibly. Candidates whose contribution for several years has been direction and oversight alone are not a fit for this role.
Evidence of Personal Contribution:
Able to describe what they personally did on the work they cite, distinct from what their team or their organisation did. Answers that stay in collective terms when a personal example is asked for will be treated as an absence of evidence.
Experience running the full people cycle for a technical team: objectives, performance assessment, difficult feedback, development and progression, including managing through managers. Evidence of retaining and developing technical staff, supported by specific examples.
Working Without a Large Support Structure:
Experience delivering in a small team without dedicated support functions to fall back on. Candidates whose experience sits entirely within very large organisations should be able to show where they have worked at this scale.
Forming an Existing Team:
Experience taking on a team selected by others, establishing how it works and getting output from it early.
Enough depth in LLM and agent architectures, retrieval, evaluation, integration and MLOps to direct and challenge engineering managers on design decisions.
Identifying AI Opportunities:
Experience working with a business function that does not yet know where AI applies, identifying where it does, and securing sponsor ownership. This includes internal corporate functions, not only client-facing ones.
Professional Services Fluency:
Able to talk credibly about how an advisory or professional services business runs: engagement teams, utilisation, the shape of a deal cycle, partner economics, where margin is won and lost, and what AI can and cannot change about any of it. This is fluency, which can be acquired outside the sector, and is tested at interview.
Structured Approach to Opportunity Identification:
Has a method for working out where AI belongs in a business, applied before any use case is named. Candidates whose approach begins and ends with a list of use cases do not meet this criterion.
Technology Stack Judgement:
Able to describe, and defend, an AI and technology stack suited to a professional services firm: models, hosting, data handling, retrieval, integration with the firm’s systems, and evaluation. The answer must fit an advisory business with confidentiality constraints, not a product organisation.
Value Realisation, Adoption and Measurement:
Has baselined, measured and reported the benefit of what they built, and tracked adoption separately from benefit. Able to describe how this was governed across a portfolio, not just on a single project.
Connecting Experience to This Firm:
Able to say what they would do here specifically, and why they are the right person for this role rather than for an AI leadership role in general. A recital of past achievements without that translation does not meet this criterion.
Data and Confidentiality Discipline:
Practical experience handling confidential client data under formal controls. Working knowledge of at least one of GDPR, UK GDPR, DPDP 2023 or POPIA, and the ability to apply the others.
Executive Communication, Presence and Constructive Challenge:
Answers the question that was asked, directly and briefly, and can hold a position under challenge. Credible in front of partners and service line leaders without support. Puts forward a better alternative when they disagree with something. Indirect answers that circle the question will be treated as a failure of this criterion, not as a style difference.
Remote and Multi-Country Delivery:
Experience being accountable for delivery to a manager in another country and time zone, serving stakeholders in several locations.