Questions people ask.

What does it cost?

A system built on our platform, hosted and supported, starts at £595 a month. Most companies we work with are five to twenty people and land between £595 and £1,200. Work that does not fit the platform is a genuine build and starts at £1,950.

We do not price per seat. The engine is already built and hardened, so what moves the price is how far your work sits from what it already does, not how many people log in. Hiring somebody should not make your software cost more, and with us it does not.

What moves the number is how much the system replaces, how many exceptions it has to handle, how much AI is involved, how many other systems it has to talk to, and how badly the company is stuck if it stops. You get a figure after one conversation, not after a discovery invoice.

Do I pay anything before I see it?

No. We build the opening screens first, in your branding, using your words for things, with demo data in them. You look at it, tell us where it is wrong, and decide.

The monthly only starts if you want the real one built. If you look at it and it is not for you, that is a perfectly good outcome and it has cost you a couple of conversations.

How long until we are actually using it?

Three to six months for a full system, depending on how much of one there is. Some are quicker: if your work looks like something we have built before, a lot of the engine already exists and we are building your screens on top of it rather than starting from an empty folder.

You are not waiting in silence. Every fortnight you get told exactly where it is up to.

What happens when the business changes?

Two different things, and we treat them differently.

Anything broken is fixed, and so is anything awkward. If a screen is slow, if something is hard to find, if the system makes you work around it, that is our error and we fix it at no cost. Improvements that make the platform better for everyone are the same. That is the job, not a change request.

New features are quoted before we start. If you want the system to do something it was not built to do, you get a day estimate and a price, and nothing begins until you have said yes. While we are still building, before you go live, changes are simply part of the build.

This is the part off-the-shelf cannot do. When your process changes, a product built for thousands of companies cannot follow you, so you work around it. Yours can just change.

Who owns it?

You own your data outright, all of it, and you can have a copy in a standard format whenever you ask. Not on request, not after a conversation about it.

The system itself runs on our platform, licensed to you for as long as you are with us. That is what lets us build yours in months rather than years and keep improving it for one fee: the engine underneath is shared and hardened by every company using it, while the part you see is built for you.

What if we want to leave?

You take your data and go. We will export it in a format your next system can read, and we would rather do that properly than make it difficult.

How long you are tied in depends on how you paid. If you paid nothing up front, there is an initial term agreed before we start, usually 24 to 48 months. It exists because we build the whole thing before you have paid for any of it, so the cost of the build is recovered over the term rather than charged at the start. The longer the build, and the more we put in before launch, the longer the term needs to be. After it, you are month to month and can stop whenever you like.

If you paid for the build up front instead, there is no fixed term at all. You are month to month from the first day.

Something breaks at seven in the morning. Then what?

You message us and a person answers. Not a ticket number, not a chatbot, and not an account manager who has to ask somebody else.

Most of the time we already know. Your system is checked every five minutes, day and night, and not merely for whether it is switched on. Whether it actually answers when asked. Whether the database is accepting work. Whether the encrypted drive is still mounted. Whether either disk is filling up, whether memory is running short, whether the certificate is close to expiring, and whether last night's backup genuinely happened. An alert fires the moment any of those changes and again when it clears, which is why nobody here has learned to ignore them.

A good number of faults never reach you at all. A crashed process restarts itself. One that starts leaking memory is recycled before it can take the server down with it. Certificates renew on their own. The ones that need a human are the interesting ones, which is the right way round.

When a person is involved: weekdays, 8am to 6pm. If the company cannot work, we start within the hour. Anything else gets an answer the same working day, not the next one. Out of hours a total outage still wakes us, because the alert goes to a phone rather than to an inbox.

Either way you are told what happened and when, rather than told it has been escalated. Our status page is public and it is checked from outside our own network, so it cannot sit there showing green while your system is down.

Do you guarantee uptime?

Not as a percentage, no. And be a little careful with anyone who does.

A 99.9% figure is easy to print and hard to check. The small print almost always measures it in a way that excludes the outages you would actually care about: planned maintenance, anything upstream, anything the supplier decides was your end. What you get if you ever claim against it is a credit against next month's fee, which is not the same thing as the day's work you lost.

What we do instead is publish the checks themselves, run them every five minutes, and keep a status page that is probed from outside our own network so it cannot report green while your system is down. When we have a year of real measurements behind us we will publish the figure we actually achieved. Inventing one today would make this page exactly like everybody else's.

What happens to our data if the server dies?

The whole database goes off the server twice a day, overnight and again after the crews finish, with the job photos alongside it. Those copies live with a different company in a different country to the machine they came from, so losing the server does not lose your data, and the day's sheets are off the box before anybody goes home.

Thirty days of copies are kept, so we can go back to a particular day rather than only to the most recent state. That matters more often than total loss does: the real case is somebody deleting six months of something on a Tuesday afternoon and nobody noticing until Thursday.

The honest limit: a copy is a copy, taken at a moment. If the server were destroyed outright between two of them, the entries made since the last one would be gone. In practice that is part of an afternoon rather than a day, because the second copy is taken at the end of one, but it is not nothing and we would rather say so.

If that is too much for your business, it is a solved problem rather than a limitation you have to accept. We can archive the database continuously instead, which brings the worst case down to a few minutes and lets us rewind to any moment rather than to the nearest copy. It changes the hosting arrangement rather than the system, so tell us early and it is priced in from the start.

Everything sits on encrypted disks on servers in Frankfurt, and all of it is exportable in a standard format whenever you want it, including on the way out.

You are a new company. What happens if you disappear?

Fair question, and it deserves a straight answer rather than a reassuring one.

Your data is yours. You can export all of it, whenever you like, without asking, in ordinary formats another system can read. It is not locked in a proprietary blob that only we can open. That is the protection that actually matters: if we stopped trading tomorrow you would take your data to another supplier and carry on, the same as if any software company you use went under.

What you would not automatically get is the platform itself, because your system runs on ours rather than being a copy handed over. That is the trade that lets us build yours in months instead of years and keep improving it for one fee.

If being able to run the software yourself whatever happens is a hard requirement of your business, say so at the start rather than at the end. That is the next question.

Can we buy it outright and host it ourselves?

No, and it is fairer to say so plainly than to quote a number nobody would pay.

Your system runs on our platform. That platform is the reason yours takes months rather than years, why a fix can land the same day, and why the fee is one number instead of a bill for every change. Handing over a copy would mean cutting it away from the engine it runs on, and what you would be left with is a frozen version with your name on it, no improvements, and your own servers, backups and security patching to look after. That is not a better position than the one you are already in.

What you get instead is the part that actually protects you. Your data is yours, all of it, exportable in ordinary formats whenever you want it, without asking and without a fee. If we stopped trading tomorrow you would take it to another supplier and carry on. That is the protection worth having, and you have it from the first day rather than as something you buy.

If running the software yourself whatever happens is a hard requirement of your business, tell us at the first conversation. We would rather say honestly that we are not the right supplier than sell you something that does not fit.

Can it talk to the software we already use?

Usually. Accounts packages like Xero, QuickBooks and Sage are the common ones, and anything with a proper API is normally straightforward.

Where something has no API, we will tell you that before you are paying for it rather than after.

We already have software. Is it worth changing?

Sometimes it is not, and we will tell you.

If your company works the way your current system assumes, it is doing its job and you should keep it. This is worth it when you are working around the software every day: the spreadsheet beside it, the things typed in twice, the report that gets rebuilt by hand every month. That is the cost worth measuring, not the licence fee.

Do you use AI in it?

Where it earns its place, and not where it does not. Reading and drafting are worth a model. Deciding whether a job has an engineer against it is an ordinary rule, and writing it as one is faster, cheaper and cannot invent an answer.

Most of what looks like AI in a Solva system is not a model at all, and costs nothing to run. Where a real model is plugged in, you pay for that usage at cost. We do not add a margin to it and we do not make anything on it, because a supplier quietly earning on how much you use something is a supplier with a reason not to make it efficient.

You get an estimate before it is switched on, a ceiling agreed in advance, and the provider's own usage figures next to the invoice. If it gets near the ceiling the work queues and we come and talk to you. A bill you did not see coming, for something you did not know was running, is not a thing that should be able to happen.

Where does our data live?

On servers in Frankfurt, Germany, encrypted, backed up twice a day, and not shared with anybody. We are not an advertising business and there is nothing to sell.

Germany is inside the EU, and the UK and EU each recognise the other as adequate for data protection, so there is nothing unusual about a UK company holding its data there. If your work means it has to stay physically in the UK, which happens in healthcare and parts of the public sector, tell us early. It is perfectly possible and it changes the hosting arrangement rather than the system.

You are the controller of the data inside your system. We handle it to run the thing, and that is set out in the agreement rather than left to be assumed.