Decide where to start, on paper first.
Fixed package fees with written outputs. Each stands on its own; none requires a build contract. Inclusions and timelines are proposed and will be finalized before publication.
AI Clarity Session
- Preparation and intake questionnaire
- 90-minute working session
- Opportunity memo
Additional scope quoted separately.
Discuss this package: AI Clarity SessionAI Opportunity Blueprint
- One business area
- Up to three stakeholder interviews
- Opportunity shortlist
- Readiness review
- Phased pilot plan
- Indicative cost model
Additional scope quoted separately. Standalone value; no build contract required.
Discuss this package: AI Opportunity BlueprintAI Adoption Roadmap
- Multi-team review
- Workshop
- Dependencies and governance
- Phased roadmap
- Executive readout
Final price depends on the number of teams and systems in scope.
Discuss this package: AI Adoption RoadmapThen build the thing.
Starting points for the most common engagements. Final price follows the agreed scope.
Scoped agent pilot
One workflow, integrated with your systems, evaluated against an agreed test set, deployed with approvals.
Product MVP
A focused first release. Final pricing negotiable based on agreed scope. Entry scope does not include a complex platform or an agent build.
Existing Projects
After reviewing code, architecture, dependencies, defects, tests, documentation, access and outstanding work.
Support and improvement
Defined coverage: monitoring, fixes, model and dependency updates, and a monthly improvement allowance. Not unlimited development.
Why an MVP can start lower than an agent pilot
The extra cost is the connection and verification work, not the model.
A CAD 5,000 MVP is a narrow first release of conventional software: one core journey, standard components, no external integrations. A CAD 10,000 agent pilot includes system integration, an evaluation set with quality thresholds, approval flows and monitoring.
What is not in the implementation fee
Billed separately or paid directly to the provider. We estimate them in the proposal so there are no surprises.
- Model and API usage
- Cloud infrastructure
- Telephony for voice agents
- Third-party services and licences
- Ongoing support
Twelve factors we look at in every scope
Reusable existing assets and a narrower first release reduce scope. Remediation of an existing codebase and custom requirements increase it. We walk through each factor with you before quoting.
- 01Features and workflowsHow many distinct tasks or journeys the first release must cover.
- 02Integrations and APIsWhether your systems expose usable APIs or need custom connectors.
- 03Data readinessClean, accessible data lowers cost; cleanup and access work raises it.
- 04Behaviour typeConversational, voice, and action-taking agents carry different engineering and evaluation load.
- 05Autonomy and approvalsHuman-in-the-loop steps need interface and audit work.
- 06Roles and scaleUser roles, permissions, and expected volume.
- 07Design maturityExisting designs and brand reduce UX effort.
- 08Codebase conditionTechnical debt and missing tests in existing projects add remediation.
- 09Testing and evaluationDepth of the evaluation set and quality thresholds.
- 10Deployment requirementsCloud, on-premise, compliance, and environment constraints.
- 11Consulting breadthNumber of teams, stakeholders, and business areas reviewed.
- 12Deadlines and supportCompressed timelines and extended support coverage.
Completing an existing project
Pricing is a custom quote because the cost depends on the state of the code. We review code, architecture, dependencies, defects, tests, documentation, access and outstanding work, then agree a completion plan. The review itself is scoped and priced before it starts.
About project completion →Want a number for your situation?
A fit call tells you which starting point applies. The assessment gives a preliminary route in two minutes.