Mosaic
Service

Custom Software Development & Rapid Prototyping

From mobile apps, websites and full applications to bolt-ons to existing software. Built from Dublin and London and starting with a working prototype rather than a specification.

Application source code on a developer’s screen

Web & mobile applications, built around how your business actually works

Most business software is a compromise. You buy a platform, then bend the process to fit it, or pay a consultancy to bend the platform back. Custom development inverts that: the software follows the process, and the process is usually where the advantage lives.

AI has changed the economics of this. Work that used to justify a year of engineering time can now land in weeks and months, so a whole class of problems that were correctly judged “not worth building” are worth reopening. Not everything should be custom, but the line has moved and most buy-versus-build decisions in place today were made on the old numbers.

We typically build for mid-sized and large organisations across Ireland and the UK, from our teams in Dublin, London and Hungary. Every engagement opens with a free discovery session and, wherever the idea can be tested, a working prototype before anyone commits to a full build.

Two ways we build

Replace the platform
Or build around it

Which shape fits the problem

The first question is not what to build. It is how much of the software you already run needs to change. Almost every engagement lands in one of two shapes, and choosing the wrong one is expensive in both directions: replacing a platform you should have kept, or papering over one that has genuinely run out of road.

Replace the platform

Build the application, retire the licence. The right call when the software you pay for is the process, and it does not fit it.

  • You pay for a platform and use a tenth of it. Licence cost scales with headcount while the value delivered does not.
  • The vendor’s model fights your differentiator. You are funding customisation or consultancy to work around a process that is supposed to be your edge.
  • The real work happens off-system. Spreadsheets and shared inboxes run alongside the platform, which costs time and destroys visibility into what is actually happening.
  • You have hit the product’s ceiling. Volume, data model or reporting requirements have moved past anything the vendor is going to build.

What we build: the whole application, covering data model, interface, integrations, authentication, reporting and hosting, plus the migration off the platform you are leaving and the support arrangement that follows.

Bolt on

Leave the system of record exactly where it is and build only the workflow around it. Lower risk than replacement, and frequently the better return.

  • The system of record is fine. It holds the data correctly. What is manual is the work feeding it and the work reading out of it.
  • You need one thing the vendor will never build. One screen, one report, one approval flow, or an integration between two products that do not talk to each other.
  • Replacement is a bigger risk than the problem is worth. Deep integrations, compliance history, or a platform that is genuinely good at its core job and weak only at the edges.
  • Low value to replace, high value to fix. The gap is a workflow gap, not a product gap, so closing it does not require owning the product.

What we build: a focused application sitting alongside the platform and integrating through its API. No migration, no retraining the business on something new. The platform stays the system of record; the bolt-on does the work around it.

In practice most organisations need both over time, and AI use cases in particular tend to arrive as bolt-ons, because the value is in reading, checking, drafting and routing work that already flows through systems you have no reason to replace. A discovery session is normally enough to tell which shape a given problem is.

Start with a prototype

Everything we build starts as a prototype not just a specification. It starts as something you can click: typically working software within two to four weeks, built against real data and real workflow so the feedback you get back is real too.

1

QUICK VALIDATION

Test the concept with the people who would use it, before the budget is committed. A prototype answers the questions often noone has even thought about.

2

ITERATIVE DESIGN

We show working software every week rather than status reports. Changing direction is cheap while the product is still small, and expensive later.

3

COST EFFICIENCY

Most of what derails a build gets discovered late. Surfacing it in week two instead of month six is the entire return on prototyping.

A prototype is also a legitimate place to stop. If it shows the idea does not hold, you have spent a few weeks finding that out instead of a year.

From prototype to production

We design, build and support the full application lifecycle:

  • Design & prototyping

    Wireframes, user flows and a clickable prototype, agreed with the people who will actually use the system rather than signed off in the abstract.

  • Development & testing

    Built in short cycles with code review, automated tests and a staging environment you can look at any day of the week, not only at milestones.

  • Launch & support

    Deployment, handover and documentation, then an ongoing arrangement for maintenance, monitoring and the next round of changes, or a clean handover to your own team if that is what you want.

The three phases of the development lifecycle: design and prototyping, development and testing, launch and support

What we build with

We pick the stack to fit the problem and the team who will inherit it, not the other way round. In practice that means Python with Django, or Node.js, on the back end; React and Next.js on the front end; React Native for mobile; and .NET where an existing estate calls for it.

Applications run on AWS, Microsoft Azure or Heroku, fronted by Cloudflare, with source control and CI/CD on GitHub and payments through Stripe. AI features are built on Azure OpenAI, OpenAI and Anthropic Claude models, integrated into the application rather than bolted on as a chat box.

Frequently asked questions

What is the difference between replacing a platform and bolting on to it?

Replacing means we build the application and you retire the licence. That is the right call when the software you pay for is the process, and it does not fit that process well.

Bolting on means the platform stays exactly where it is as the system of record, and we build only the workflow around it: the screen, report, approval flow or integration the vendor will never build. It is the lower-risk option and often the higher-return one.

How long does a prototype take?

Most prototypes are working software within two to four weeks, built against your real data and real workflow rather than sample content. You get something you can click and put in front of the people who would use it, not a specification to argue about.

Who owns the code and the intellectual property?

You do. Everything we build is delivered into your own repositories and your own cloud accounts, with no licence back to us and no dependency on Mosaic to keep running it.

Do you support the software after launch?

Yes. Launch includes deployment, handover and documentation, and most clients continue with an ongoing arrangement covering maintenance, monitoring and the next round of changes. If you would rather your own team took it on, the handover is built to make that possible.

Can you work alongside our in-house development team?

Regularly. We can lead delivery with your team embedded, contribute to a team you already run, or hand a prototype over for your developers to take to production. Where the need is extra engineering capacity rather than a defined build, our developer outsourcing service is usually the better fit.

Do you only work with companies in Dublin?

No. We are based in Dublin and London with team members in Hungary and the Middle East, and we work with organisations across Ireland, the UK and Europe. Discovery and key milestones are usually in person where geography allows; the rest of delivery is remote.

Not sure whether to replace it or build around it

That is what the discovery session is for. Bring the process that is costing you the most and we will tell you which shape the fix takes, free of charge.