You start with a defined pilot.You only scale if it convinces you.
A fixed project on your own history. If the model works, we implement it; if not, you stop there. And the subscription is optional after the first year.
Pilot
One use case, validated on your history. 3–4 weeks. One-off payment. Deducted from the implementation if you continue.
- Cleaning and structuring of your Excel
- A trained and validated model
- Trial deployment on your network
- Accuracy report and comparable cases
- You keep the model, whether you continue or not
Implementation
Full rollout on your server, with your processes and your team. Minus the pilot already paid.
- Several models in production
- On-premise deployment on your network
- The Excel you already use as the data source
- Training for the technical team
- Backup = a single file
Subscription
Only if you want to keep the model alive. Start it, stop it or resume it whenever you like.
- We retrain with your new data
- We check that your models keep their accuracy
- Software updates
- Enquiries and direct support
- No per-user fee, whoever uses it
Prices exclude VAT. Requires your own server (an office PC is enough to start). Prediction quality depends on the quality and quantity of your history. Every prediction comes with its margin of error and supporting cases, never a bare number. The system proposes; the quoting decision is always yours. And the fee, when there is one, is per installation, not per user.
Today's alternativedoesn't scale
You aren't up against other software. You're up against Excel and one person's experience, and that person isn't always available.
| Excel + experience | Helenus | |
|---|---|---|
| Time per estimate | Hours or days | Minutes |
| Margin of error | Not documented | Explicit, always visible |
| You know when not to trust it | Never | Range + confidence when the data falls short |
| Depends on one person | Yes: holidays, sick leave, departures | No, the model stays in the company |
| Data exposure | Loose files, no control | Everything on your network, verifiable |
| Scales with more cases | Slower and slower | Improves with more history |
Before you ask
What if the margin of error comes out high on my data?
We'll tell you, and that's the important part. If your history can't support predicting an exact amount, the system says so plainly and recommends moving to prediction by ranges, which is reliable and actionable. An honest 50% margin is more useful than an invented 5% that will lose you money on the first real quote.
Is the monthly payment mandatory?
If you stop at the pilot, no: it gives you a model that is yours and keeps working on your network without paying anything more. If you implement, the subscription is required in the first year and optional after that. It lets you keep retraining with new data, monitor that it keeps its accuracy, receive updates and have support for enquiries.
Do I need my own server or special infrastructure?
No. A normal office PC is enough to start. If you later want several people to access it at once, it's deployed on a simple local server, but never in a third party's cloud.
What if my history is only 20–30 cases?
That's the usual scenario in smaller companies, and the one the system is designed for: we use models that hold up on small histories rather than those that need thousands of cases. And cross-validation instead of a single split, which gives an unreliable figure with little data.
Can I try it before committing to anything?
Yes. That's exactly what the pilot is: a defined 3–4 week project on your own history, with no lock-in afterwards. If the model doesn't reach an acceptable margin of error, you stop there.
What happens to my data during the pilot?
It's processed on your machine or in an environment agreed with you, never on a server of ours shared with other customers. When the pilot ends, the model and the data stay wherever you decide.
Is it only for quoting?
No: the same pipeline predicts any variable you set as the target: cycle time, rejection risk, mould life. Quoting is usually the first case because it's the easiest to justify financially.
Sounds like a fit?Tell us about your case.
We'll show you how it would look with your own history. We reply within 24–48 hours.