redplumm

Home › Desktop, agents and integration

Desktop, agents and integration

Windows services, background agents and the connecting work that makes a new system talk to the accounting, payroll and stock software you already pay for.

The work nobody demos

Integration is the least glamorous part of a software project and routinely the part that decides whether it succeeds. A new system that cannot see your stock levels, cannot push an invoice into your accounts package and cannot read your payroll is not a system — it is a fourth place to type things, and your team will quietly stop using it by March.

We do this work as a matter of course, and we also do it on its own, for systems we did not build. If the problem is that two things you already own refuse to speak to each other, that is a perfectly good reason to call us.

What usually needs connecting

Accounting and finance, so invoices, credit notes and payments stop being rekeyed. Payroll and time capture, so hours recorded in the field arrive without a spreadsheet in between. Stock, ordering and point of sale. CRM and email. Document stores. Supplier and customer systems that expect you to meet them in their format, on their schedule.

Some of these offer a clean modern API. Many do not. A great deal of real integration work is a nightly CSV on an FTP server, a database only reachable from one machine in the office, a fixed-width file whose layout exists in one person's memory, or an application with no interface at all beyond the one a human clicks. All of those are workable. None of them are tidy, and we would rather be honest about that in advance than discover it in week six.

Windows services and background agents

Where the work has to run on a schedule, unattended, inside your network, we build it as a proper Windows service in C# and .NET rather than a script in Task Scheduler that silently stops working one weekend and is discovered three weeks later.

That means it starts with the machine, recovers from failure, writes a log somebody can actually read, and tells you when it has a problem instead of failing in silence. Where something needs to run inside a browser against a system with no API, a purpose-built extension is sometimes the cleanest route, and we build those too.

Legacy systems are not a barrier

We work with plenty of software that is older than the business case for replacing it. A twenty-year-old system that runs the operation reliably is not a failure, and ripping it out is rarely the cheapest answer. More often the right move is to leave it alone and build around it — put a modern interface in front of it, sync what is needed out of it, and let it keep doing the one job it does well.

Replacing it, if it comes to that, is then a decision you make deliberately rather than one forced on you by an integration that could not be made to work.

Starting small is fine

Integration is one of the few things we are happy to take on as a contained piece of work rather than a full engagement. If one connection would remove a daily hour of retyping, that is worth doing on its own merits.

The first conversation is free and the useful part is usually a list: what systems you run, what moves between them today, and who is currently moving it by hand. Bring that list.

Related: custom software development, web platforms and portals.

Start with a conversation

The first meeting is free and its job is working out whether this is worth building at all. If it is, you will see a working prototype before any long-term money is discussed.

Book a slot   or call 020 4652 7504

Redplumm International Limited · United Kingdom, with a remote office in Johannesburg, South Africa