Data & AI Consulting: Scoping, Strategy, Governance and Change

Engagements, KPI trees, business cases, UK GDPR, DPIAs and adoption

For analysts moving into client-facing work and consultants new to data and AI: scope an engagement, build a KPI-aligned business case, screen it for data protection under UK and EU GDPR, and plan adoption so the work is used.

Intermediate · 4 hours · 34 lessons · 51 labs

Who it is for

Intermediate level.

Before you start:

Labs

51 labs, 37 of them graded: you write the code, it runs, and graded labs check the result. The outline lists each lesson's labs.

What is free

Start free: First module free with an account. Pro: Every course, lab, handbook, practice exam and certificate.

Course outline

Start Here

  1. The Dashboard Nobody Used (1 lab): You can explain why a technically sound analytics product goes unused, and name the four questions a consultant asks before anything is built.
  2. What Data And AI Consultants Do (1 lab): You can describe the roles on a data and AI engagement, what a client buys from a consultant, and where analysis ends and advice begins.
  3. The Engagement Life Cycle, From Proposal To Handover (1 lab): You can lay out the phases of an engagement (proposal, mobilise, discover, design, deliver, hand over), the decision at the end of each, and what goes wrong when one is skipped.

Client Leadership

  1. The Trusted Adviser: Credibility, Reliability, Independence (1 lab): You can use the trust equation to diagnose a client relationship and choose words and actions that build credibility, reliability and intimacy while lowering self-orientation.
  2. Ethics, Codes Of Practice And Conflicts Of Interest (1 lab): You can apply a professional code of conduct to consulting dilemmas, spot a conflict of interest, and record and manage it.
  3. Proposals And Statements Of Work (3 labs): You can write a proposal and statement of work whose objectives, scope, exclusions, deliverables, acceptance, assumptions, fees and change control leave no room for argument later.

Scoping And Requirements

  1. Stakeholder Mapping (1 lab): You can map stakeholders by power and interest, choose how to engage each group, and find the person whose decision the work depends on.
  2. Discovery Interviews (3 labs): You can plan and run a discovery interview with open, non-leading questions, follow the evidence, and capture what you heard without bending it.
  3. Problem Statements And Issue Trees (1 lab): You can turn a vague request into a problem statement with a decision, a measure and a deadline, and break it into a mutually exclusive, collectively exhaustive issue tree.
  4. Requirements, Acceptance Criteria And MoSCoW (1 lab): You can write requirements with testable acceptance criteria and prioritise them with MoSCoW so the Must haves fit the time box.
  5. Success Measures And The Scope Baseline (1 lab): You can agree baseline and target measures before work starts, freeze a scope baseline, and run change control against it.

Data And AI Strategy Aligned To KPIs

  1. From Goals To KPI Trees (1 lab): You can turn a business goal into a KPI tree whose branches multiply or add to the top measure, and use it to find where an analytics product can move the number.
  2. Assessing Data Maturity (2 labs): You can run a data maturity assessment with a published framework, read the weakest theme, and say what it means for what to build first.
  3. What AI Solutions Are Made Of (1 lab): You can describe the main kinds of AI solution a client may be offered (predictive models, generative model applications, retrieval over documents, agents that call tools) and what each needs in data, cost, skills and risk.
  4. Assessing AI Fit (1 lab): You can decide whether a problem needs AI, a simpler rule or report, or no technology at all, using explicit criteria.
  5. Prioritising Use Cases: Value Against Feasibility (1 lab): You can score use cases on value and feasibility with weights the client agrees, rank them, and test whether the ranking survives a change of weights.
  6. The Business Case: Costs, Benefits, NPV And Payback (1 lab): You can build a simple business case with one-off and running costs, benefits that ramp up, net present value at a stated discount rate and payback period, and say what it is most sensitive to.

Data Governance

  1. Owners, Stewards And Custodians (1 lab): You can assign data ownership, stewardship and custodianship for a data domain, and check a RACI has exactly one accountable person per decision.
  2. Data-Quality Dimensions And Rules (1 lab): You can name the six data-quality dimensions of the UK Government Data Quality Framework, write a testable rule for each, and report quality as pass rates against thresholds a business owner can act on.
  3. Metadata, Lineage And Catalogues (1 lab): You can describe business, technical and operational metadata, trace lineage from source to report, and use it to find what breaks when a source changes.
  4. Operating Models And Policies (3 labs): You can compare centralised, federated and hybrid data governance operating models, and write a policy, standard and procedure at the right level.

Privacy And Compliance

  1. The Seven UK GDPR Principles And The EU GDPR (1 lab): You can state the seven data protection principles, apply them to an analytics plan, and say where the UK and EU regimes now differ.
  2. Lawful Bases, Including Recognised Legitimate Interests (2 labs): You can choose and record a lawful basis for an analytics purpose, including when a legitimate interests assessment is needed and what the new recognised legitimate interests are.
  3. DPIAs: When And How (1 lab): You can screen a project for whether a DPIA is required, using Article 35(3) and the ICO's list of processing likely to result in high risk, and outline the DPIA's contents.
  4. Rights, Retention And Records (2 labs): You can handle the individual rights an analytics team meets, set and enforce retention, and keep a record of processing.
  5. Pseudonymisation And Anonymisation (2 labs): You can pseudonymise identifiers with a keyed hash, explain why pseudonymised data is still personal data, and test a release for k-anonymity.
  6. The EU AI Act For Analytics Teams (1 lab): You can place an analytics or AI use case in the AI Act's risk tiers, name the dates that apply, and say what a UK team serving EU customers must check.
  7. Guardrails, Privacy And Governance (4 labs): You can check what goes into and out of a model, redact personal data, render model output safely, and name the duties NIST AI 600-1 and the EU AI Act place on an AI feature.

Change Management And Adoption

  1. Why Analytics Fails To Land (1 lab): You can diagnose why an analytics product is not used, from an adoption funnel and from the people side, and name the fix for each stage.
  2. Change Models: Kotter's 8 Steps And Prosci ADKAR (1 lab): You can use Kotter's 8 steps to plan an organisation-level change and Prosci's ADKAR to find where individuals are stuck.
  3. Designing Adoption: Training, Champions And Measures (1 lab): You can design an adoption plan with role-based training, a champion network and adoption measures with targets.
  4. Handover And Sustaining Value (1 lab): You can plan a handover that leaves the client owning the product, and track benefits after the consultants leave.

Capstone

  1. Capstone: Scoping Memo And Roadmap (3 labs): You can write a one-page scoping memo for a client and sequence a roadmap that respects dependencies.
  2. Capstone: DPIA Screening And A Steering-Committee Update (3 labs): You can screen the capstone project for a DPIA, and brief a steering committee in a way that gets a decision.