Choose a package
Each package is a complete engagement with defined deliverables. If your project falls between two tiers, we will quote a custom price after a free 30-minute scoping call.
Strategy
Best for organisations exploring where AI fits into their operations.
- Three-day on-site or remote workshop
- Data-asset inventory
- Prioritised use-case roadmap
- Half-day AI literacy training
- Written report within 5 working days
Build
A single predictive model or NLP pipeline, from data audit to production deployment.
- Data audit and cleaning (2 weeks)
- Model training with 3+ algorithm comparison
- Bias audit and accuracy report
- Production API or batch deployment
- 6 months monitoring included
- Full source code and documentation
Scale
Ongoing AI support for companies running multiple models or expanding automation.
- Dedicated senior engineer (20 hrs/month)
- Model retraining and drift monitoring
- Monthly performance report
- Priority bug fixes (4-hour SLA)
- Quarterly strategy review call
Feature comparison
A side-by-side view so you can see exactly what each tier includes.
| Feature | Strategy | Build | Scale |
|---|---|---|---|
| Data audit | Overview only | Full audit + cleaning | Ongoing |
| Model training | — | 1 model, 3+ algorithms | Retraining existing models |
| Deployment | — | API or batch | Maintained |
| Monitoring | — | 6 months | Continuous |
| Source code handover | — | Yes | Yes |
| Bias audit | — | Yes | Quarterly |
| Dedicated engineer | Workshop facilitator | Project lead | 20 hrs/month |
| Support SLA | Email (2 days) | Email (1 day) | 4-hour response |
| Minimum commitment | One-off | One-off | 3 months |
Common questions about pricing
Three factors: the number of data sources we need to integrate, the volume of data (which affects compute costs during training), and whether the model requires a real-time API or a nightly batch run. A single-table churn model for 50,000 records sits at the lower end. A multi-source NLP pipeline processing 500,000 documents with a sub-second API response requirement sits at the upper end. We provide a fixed quote after the scoping call, not a vague estimate.
No. The first call is 30 minutes, free, and has no obligation. We use it to understand your problem, ask about your data, and decide whether we are the right fit. If we are not, we will say so and suggest alternatives where we can.
We define a minimum accuracy threshold in the scope document before work begins. If the final model falls below that threshold after our best efforts, you do not pay the final milestone (typically 30% of the project fee). We eat that cost. This has happened twice in 42 projects; in both cases the root cause was insufficient training data, which we flagged as a risk during discovery.
Yes, and many clients do. The Build package includes six months of monitoring. Before that period ends, we review whether ongoing support makes sense. If you move to Scale, the transition is seamless because the same engineer who built the model continues to maintain it.
Our prices cover our labour: engineering, data science, project management and documentation. Cloud compute and storage (AWS, Azure or GCP) are billed separately at cost, with no markup. We estimate these costs in the scope document so there are no surprises. For a typical Build project, cloud costs during training run between £200 and £800; ongoing inference hosting is usually £40 to £150 per month depending on traffic.
Need something different?
If none of these packages fits your situation, get in touch anyway. We have structured custom engagements for clients who needed a proof-of-concept in two weeks, for teams that wanted us to train their internal data scientists rather than build the model ourselves, and for organisations that required on-premise deployment with no cloud dependency at all.
Call us on +44 7772 427025 or email [email protected]. We respond within one working day.