There are lots of things in your call centre that you may put together yourself rather than buying in software or hardware: reports, scripting software, intranet sites and work force management to name but a few.

Often the choice to do this depends on the size of the centre. For example, It’s far more common to see centres with less than 100 seats using Excel for workforce management.

In fact, up until recently there was only one piece of kit that you couldn’t put together yourself: the telephone system.

The DIY Switch

Amazingly, you can now build a telephone system yourself from separate components. Granted, it’s still not cheap - but you can save money. I could reel off a list of advantages and disadvantages as long as my arm - but I charge good money for that - so I’ll just say that the important considerations are stability and supportability, both of which could be better or worse with a DIY system instead of something off the shelf.

Assuming that DIY would be suitable for your company, how would you go about the task?

Hardware

First, you’ll need a server.

You can skip this next bit if you’re not techie inclined…

Telephony components are available in two main flavours nowadays, known as H.100 and H.110. H.100 are PCI boards, which mean they’ll fit in any old bog standard server. Because the PCI bus doesn’t provide everything a telephony connection needs all H.100 boards must be interlinked within the server by a ribbon cable. The new H.110 standard is requires a specialist server as the cards are the new cPCI format - but there’s no need to interlink and the hardware is far more robust.

…now pay attention again!

The server in theory could be any old desktop PC, but you may want to consider how stable and reliable this would be. Purchasing a high-end server such as a Westek will give you a passive backplane, multiple power supplies and better electro-magnetic shielding. All these things are going to give you a more solid platform.

To complete the hardware you simply need to install PCI or cPCI cards to provide telephony connections. You’ll need an E1 card (Available single or dual, for 30 or 60 lines) to connect to the ISDN lines. You’ll need station cards to provide a physical connection to your telephones (Usually available with 4,8 or 16 ports) and resource cards to provide signal processing (for tones and such, 1 port for each station port). Complicated as it may sound, this is no different from specifying hardware for a proprietary telephone system.

These cards are now manufactured by many people, but the two big names are Intel (previously Dialogic) and Pika.

Software

You can purchase software to run on this platform, or you could get something bespoke written.

If you chose to purchase software then be sure that it’s S.100 compliant. This means that it’ll work with the H.1*0 hardware you’ve got, and will also behave properly with other S.100 compliant software running on the same system. This means it’s possible to be running a call queuing package from one supplier and an IVR system from another on the same server. Now that does sound like fun…

Writing the software bespoke could be a massive task, especially if you want to write an inbound call handling package with IVR and call recording. But if all you want to do is launch outbound calls then there’s not so much work to do and it’s a lot more achievable. The work involved is further reduced if you use some kind of SDK such as those supplied by card manfacturers. These also give you an element of hardware abstraction - which means that if you change the cards within the system it won’t get so confused. There are higher level SDK’s available to give very quick programming and complete hardware abstraction, but these are quite pricey and it’s questionable whether this would give you any savings on purchasing the software.

The future
This is a very interesting part of the telephony market and I believe we’ll see it expand over the next few years. System manufacturers will see the benefits of using standard hardware, and less and less companies will be interested in creating their own. I think this will happen in the same way that the advent of the IBM PC resulted in nearly every computer being a clone of this model.

In the long term this means better and cheaper systems as energy will be focussed on innovation rather than re-inventing the wheel.

If you’re going to give it a try…
…then think about getting someone in to help you:

Regards,
DB