mainframe chatbots

For years now, major enterprises – including banks, airlines and insurance companies the world over – have remained faithful to the mainframe to run their core business systems. No surprise, as we all know it remains one of the most dependable, secure and hard-working systems on the planet. 

Sadly, that still might not be enough to guarantee the longer term future of the platform, as IT leaders continue to worry about the dearth of new talent ready to step into the shoes of the current crop of mainframe specialists as they retire. Could AI-driven chatbots be part of the answer by helping to make the mainframe more accessible to newbies? I certainly believe so.

We expect tech to be easy

Today, the ongoing consumerisation of IT means we all – mainframers included – demand technology that is easy to use. We’re accustomed to apps with simple, graphical interfaces and we expect to tell the likes of Siri and Alexa what to do and for it to just happen. Compared with this type of user experience, the mainframe’s ‘green screen’ command line interface feels like technology from a bygone era.

Younger people especially view the z/OS platform as old-fashioned, uncool and clunky. As digital natives, they expect technology to be intuitive: something they should be able to figure out for themselves. The mainframe just isn’t like that.

The steep mainframe learning curve

Understanding Big Iron takes time and effort and there is much to learn before you can feel comfortable and productive. It’s a challenge for today’s computer science graduates as most will not have come face to face with a z/OS system while at university or college.

At the same time, there are fewer experienced mainframe specialists around to help train the next generation. Those who haven’t already retired have very little time to help, simply because they’re so busy now, due to the mainframe skills shortage.

Many enterprise IT chiefs feel exposed because the core systems that still run on the mainframe were developed back in the 70s and 80s in languages they are unfamiliar with.  Many CIOs have little or no direct experience of working with the mainframe, and the information that comes back from their mainframe specialists is so technical that it makes little sense to them. IT execs need to plan for the future and you can’t blame them for wondering if it’s time for a change.

Solving the IBM Z user experience challenge

In a previous post I discussed how one of the mainframe’s biggest problems is poor user experience.

Some mainframe shops are trying to address this by writing ‘good looking’, user-friendly, customized interfaces that support web and mobile access for their most important mainframe applications – especially those that are accessed by customers.

Another solution is offered by modern session managers which offer web enablement and a more user-friendly front end for the mainframe. These present applications within a web browser that responds to mouse clicks and touch screens; it’s a way of making the mainframe accessible from any PC or mobile device without using expensive terminal emulators, with their footprint and administration overheads.

AI-chatbots for a user-friendly Big Iron experience

Now AI-enabled digital interfaces or chatbots have arrived and they promise to make another huge leap towards making the mainframe platform more user friendly. In fact, we may just have found the missing piece of the puzzle that solves the mainframe’s user experience problem.

AI and chatbot technology has now advanced to the point where conversational interfaces can be built into or around mainframe software to support users. They can act like digital trainers or guides to help new recruits find their way around mainframe systems, answering questions and passing on relevant technical knowledge to make novices more productive more quickly.

The application of chatbots in every realm of our lives is growing rapidly. They are used for answering financial queries, providing customer support, diagnosing medical issues, and even offering counselling. Experts are predicting that this kind of AI could be used to support continuous employee training, including the transfer of skills from one generation to the next, as experienced employees move on or retire – exactly what’s required in the mainframe space.

An AI knowledge base for mainframe newcomers

Put simply you could ‘educate’ a chatbot about your mainframe system so that it can pass on knowledge to less experienced colleagues. Imagine using IBM Watson to populate a knowledge base with IBM Redbooks, as well as comprehensive information on your mainframe applications. Now imagine being able to question a chatbot that can immediately recall any part of that knowledge base whenever you need it – or proactively offer guidance if it detects a situation where assistance is required (and even assess whether help would be welcomed, using sentiment analysis).    

The beauty of chatbots that are supported by AI and machine learning is that they can learn and get better with each interaction or conversation, detecting preferences and making recommendations based on past requests or scenarios. By interacting with mainframe veterans, chatbots offer a way for organizations to absorb and retain the vast store of knowledge their workforce has collectively amassed over the years and make it available to the next generation.

Easing mainframe support challenges

One obvious area where AI-assisted chatbots can help is in mainframe support and problem resolution. For example, the bot could flag up a problem for a support engineer or systems administrator to fix; navigate to the part of the system where the problem resides; and provide context-sensitive guidance, backed up by relevant metrics, to assist rapid analysis and diagnosis.

A bot could provide instructions on how similar problems were fixed before; or identify if there is an expert on the team who has resolved an identical problem – and bring that person into the conversation in real-time.

Increasingly you can expect to see situations where chatbots are working together – for example by escalating highly technical queries to other specialist bots that have deeper expertise in specific technologies, both on the mainframe and other platforms. There is also the possibility of ‘mixed reality’ conversations between teams of people and teams of bots as they try to resolve particularly complex problems. In the future the boundaries between humans and AI may start to blur as it becomes harder to tell who is real and who is a bot!

It’s happening already

How close are we to these kinds of scenarios being part of our everyday technical mainframe environment? Closer than you might think. Take the example of Nokia, which has introduced a digital assistant that is trained to assist technicians with telco-specific tasks, including diagnosing and analyzing the root causes of network faults and helping to “build their network knowledge through human-like interaction”. As part of UNICOM’s mainframe modernization development initiative I am working with our teams to ensure we also take advantage of AI-assisted chatbots, both internally and within our mainframe offerings.

Analysts predict that eight billion voice assistants will be in use by 2023 as the Internet of Things becomes a reality. We’ll be talking to our phones, fridges, cars and all manner of devices in our homes and workplaces. We’re going to be very used to communicating with ‘things’ with our voices, so it seems only natural that voice-enabled bots will become an important way to interact with the mainframe, too.

Whether they respond to voice, online chat or both, the digital assistants of the future will create a more supportive work environment that facilitates learning, and enables a faster time to productivity for up-and-coming mainframe developers and operations staff. They are all set to reduce the effort and costs involved in managing heritage IT systems and help us to repurpose the mainframe for a new generation.

Keith Banham has worked in IT for over 35 years and is the mainframe R&D manager at Macro 4, a division of UNICOM Global. Keith started as an Assembler programmer at a major bank and during his 30+ years at Macro 4 he has worked on many of the company’s solutions for application lifecycle management, application performance management, document management and session management. He is responsible for driving the modernization of these solutions by building web, Eclipse and mobile interfaces, and architecting cross-platform solutions utilizing UNICOM’s open systems and IBM i capabilities.

Leave a Reply

Your email address will not be published. Required fields are marked *