Customer Conversations

Start with the work customers need to change.

A campaign must meet the client's brief. An investigation needs defensible findings. An engineer must explain a solution in its business context. These conversations show why the new workforce needs more than access to AI tools: people and agents must work together through clear decisions and handoffs.

Anonymized summaries of customer and partner conversations. Buyer needs are recorded requests; the rehearsal and evidence below describe preparation to agree, not measured business results or a completed deployment.

From a buyer need to shared-work readiness

Understand the workflow and its existing tools. Use Task Intelligence to examine responsibilities, then agree simulations that reveal what happens in normal work, exceptions and handoffs. Assess people, evaluate the selected agents and let operating owners accept the shared workflow against separate criteria.

Media business

Campaign brief, creative production and client review

Buyer request: improve an existing workflow

What the buyer needs

Sales and creative teams already use AI, yet client feedback drives repeated changes. The need is consistent quality and authenticity across ideas, pitch materials and the creative work that follows an order.

What people own
Sales interprets the client brief. Creative specialists judge relevance, voice and authenticity. A named owner approves the work presented to the client.
Agent contribution to explore
Explore support for research, draft concepts and creative variants. Generated material still needs review against the brief and the agreed quality standard.
What to rehearse
Give sales, creative and the AI contributor the same brief. Introduce an incorrect pronunciation or a client change, then practise deciding what to revise and who approves it.
Separate evidence to review
Review people's quality judgments, the selected agent's output against the brief, and whether the revision handoff preserves the client's requirements.

Recorded scope: Proposed preparation. A request for a more consistent production process is not evidence that an autonomous creative factory has been delivered.

Professional services team

Investigation evidence, findings and report review

Buyer request: preparation in approved existing tools

What the buyer needs

Investigators want to use their existing AI assistant responsibly after case data is available. They need to detect unsupported statements, verify policy references and preserve accountability for their findings.

What people own
Investigators verify evidence and distinguish a finding from a conclusion. The engagement owner reviews the recommendation; the client decides the next action.
Agent contribution to explore
Explore assistance with retrieving evidence and drafting findings. Every material statement must be checked against the approved case sources and relevant policy.
What to rehearse
Use an approved or synthetic case pack in the team's existing tools. Introduce a plausible statement without supporting evidence and practise correcting or escalating it before report review.
Separate evidence to review
Review human verification and escalation, agent grounding against the case pack, and whether report handoffs preserve uncertainty and decision ownership.

Recorded scope: Proposed simulation scope. Existing tools can support practice without a separate lab; source grounding does not guarantee hallucination-free output.

Learning platform and services partner

Technical practice, FDE discovery and demonstration

Reported lab-session use; expanded technical and FDE scope under discussion

What the buyer needs

Technical learners need useful practice environments, an AI coding assistant and cloud access. FDE preparation also needs the stakeholder conversation that turns a business need into a bounded build and a demonstration.

What people own
The engineer asks discovery questions, checks the problem boundary, reviews code and explains the solution. The partner agrees learning objectives, access and delivery ownership.
Agent contribution to explore
An AI-simulated stakeholder can support discovery practice. A coding agent can contribute to the build. These play different roles and require different checks.
What to rehearse
Start with an ambiguous stakeholder request, build in an agreed environment and demonstrate against the agreed problem. Introduce an integration constraint that requires the engineer to renegotiate scope.
Separate evidence to review
Review human discovery and technical judgment, coding-agent behavior in the chosen environment, and whether the demonstration and handoff match the agreed problem.

Recorded scope: A lab-based session was reported as used. New environment variants and FDE expansion require their own scope, tool and licence review.

Enterprise technology services team

Engineering preparation inside an industry workflow

Buyer proposal: co-design an industry scenario

What the buyer needs

Technical labs alone do not explain the business. The request is for industry-specific scenarios in which engineers can understand transactions, systems and decisions before proposing a change.

What people own
Domain specialists define a bounded business scenario. Engineers explain the change and its trade-offs. Business and architecture owners review the proposed solution.
Agent contribution to explore
Explore agent contributions within the selected workflow, with explicit inputs, permissions and escalation points rather than assuming a complete industry can be automated.
What to rehearse
Choose one industry workflow with the domain team. Introduce a missing input or conflicting requirement, then rehearse how the engineer, agent and business owner resolve it.
Separate evidence to review
Review human understanding of the business, agent performance on the agreed step, and the shared response to the exception.

Recorded scope: Co-design was discussed. A supplied business blueprint is a starting point for scoping; it does not establish a complete simulated industry landscape already delivered.

Enterprise engineering programme

Enterprise engineering cohorts and capstone handoff

Recorded environment requirements; delivery scope to agree

What the buyer needs

Engineers need practical retrieval and agent-building projects using existing licences, suitable cloud services and code they can retain. Discovery-conversation practice is a separate option to evaluate.

What people own
Engineers define and test the build, review the evidence and explain its limits. Internal SMEs agree criteria and the receiving team reviews the code and handoff.
Agent contribution to explore
Coding and retrieval agents can contribute within the approved environment. The project team defines what they may access and how their outputs will be checked.
What to rehearse
Use a bounded capstone with approved practice data. Introduce a permission or retrieval failure, then practise diagnosing it and handing over the code, assumptions and unresolved limits.
Separate evidence to review
Review human implementation and explanation, the selected agents' behavior on test cases, and whether the receiving team can understand the handoff.

Recorded scope: A recorded request, not a completed programme. Agree access, credentials, licences, retention, support and assessment ownership before delivery.

Bring the workflow you already have.

Start with your people, agents, approved tools and the decisions that matter. Agree the preparation you need. The AI Bootcamp is a 6 to 8 week engagement; rollout and continuing support are separately scoped.

Looking for market studies? Explore the research behind the framework.