The helpdesk is caught between a rock and a hard place. On the one hand, there's the desire of end-users to have their hands held. On the other, there's the drive to cut support costs. Technology has been deployed increasingly in an effort to ease the pressure, with self-healing and self-help reducing the number of helpdesk calls per user, although this is often offset by the increasing complexity of IT systems and software products.
This trend will continue, given that even today most helpdesk calls involve basic questions that have either been answered before, or involve straightforward information that could be supplied online.
According to Christopher Harget, product manager at helpdesk systems supplier Remedy, 30% of calls to helpdesks are status checks from end-users wanting to find out how soon a problem they reported earlier will be resolved. And of the remaining calls reporting a new problem, 80% involve questions that have been resolved satisfactorily before, says Don Ross, director of business development for Knowledge Management Software.
Combining these two figures reveals that only 14% of all helpdesk calls are new problems that require serious attention. The remaining 86% could all be resolved automatically without human intervention, via Web-based interfaces.
Even some of the new-problem calls could also be avoided through a combination of self-healing capabilities, and better training of the end-users in the first place so that they are less likely to run into difficulties. Some of these capabilities are already in place, thereby reducing helpdesk calls, but there is a lot of potential yet to be tapped. There is also scope for reducing helpdesk calls by maintaining systems better so that end-users are less likely to have problems, and integrating technical support with change management so that any bugs identified in the process can be fixed.
A point emphasised by many helpdesk suppliers is that cultural issues will remain as relevant as ever for future helpdesks. There will still be a substantial minority of end-users, including staff and external customers of Internet services, who insist on dealing with a human operator, and perhaps cannot cope with the automated system. So while it may be possible in theory to eliminate more than 80% of the calls currently made to helpdesks, in practice the saving will be substantially less. But automation can also cut costs by assisting first line support staff, reducing both the level of skill they need and the time they take to answer users' questions.
The original move from on-site to telephone support was made largely to cut costs, given that many users would still prefer the former.
But even telephone support, assisted by online systems, is relatively expensive and, says Harget, there will be a big drive towards use of instant messaging for communication between end-users or customers and helpdesk staff within the next four years. This, he says, will reduce the time wasted by helpdesk operators while end-users fumble with their keyboards trying to follow instructions. It could also reduce the frustration of end-users caught in queuing systems waiting for the helpdesk to reply.
"With instant messaging, providing it is tightly integrated into the helpdesk, an operator could be carrying on conversations with several end-users at once with none of them noticing anyone else was involved," he says. The use of text-based conversations will also make it easier to keep records of each problem for access by the helpdesk.
It might seem that most of the anticipated developments are aimed at cost cutting but there are some that should increase the quality of life for end-users.
The growing availability of broadband communications within the next four years will make it feasible to deliver mixed media, interactive helpdesk systems, allowing, for example, the use of video clips on demand to illustrate particular sequences of action. There are already some tasters of what is to come, such as US company
SeeItFirst.com's offering which has solved the resolution problems that blight many video-based training or self-help systems over lower bandwidth connections. This makes it impossible to freeze a frame and zoom in to examine a particular image in detail and at leisure. SeeItFirst.com allows users watching a video sequence to click on a frame and freeze it. The system will then access the frame from a server holding a much higher resolution version of the video, grab the requested frame, and download it to the end-user as a JPeg file. In that way, the user can obtain a single frame at a much higher resolution than can be delivered as part of a continuous video stream at the available bandwidth. But when broadband communications become available, it will be possible to provide whole video sequences at a high resolution. "The potential for broadband communications to improve helpdesks is huge," says Harget.
Knowledge management software also has potential for enabling end-users to access the help they need when they need it. This will be combined with personalisation services tailoring the helpdesk system to individual needs. It will also enable the helpdesk system to answer queries made in natural language, by identifying the nature of the question from the combination of words.
According to Ross, natural language combined with neural network technologies will enable textual queries to be matched with the solution in a knowledge base. The term "neural network" may have been overused in an attempt to associate what is essentially intelligent adaptive pattern matching with the human brain for marketing purposes, but it can be very successful.
There will also be progress in adaptive help, attempting to provide assistance automatically as end-users encounter difficulties. Again, this would tie in with a knowledge base of solutions, attempting to identify the nature of a problem before the end-user has even got round to reporting it. But care must be taken here not to be overbearing, says Ian Robinson, account services and operations director at IT consultants CWB.
As he points out, rudimentary versions of such adaptive or self-healing functions are enshrined in software products such as Microsoft Word. "People often find this annoying, and many complain when the Microsoft winking paperclip appears and claims to know what the end-user is trying to do."
Another significant development over the next few years, and already under way, is a growing convergence between online help and e-learning, with each incorporating many of the same interactive and multimedia features.
It is clear then that many of the features of the future helpdesk are already in place and need to mature and be accepted by end-users.
But as Robinson says, an ideal world without any IT problems is unlikely so long as humans continue to be involved in the building of systems.
Helpdesks of the future