IBM’s Koala – Screencasting on Steroids

You must have all seen screencasts – a stream of screenshots with annotations of a session with a software application. This is the technique we use to explain how to use a software system. The person learning the new system must painstakingly go through the screencast, do the same actions on the system and learn to use the system step by step.

I have wondered for a long time, why we couldn’t create the equivalent of training wheels (which kids use to learn cycling) for learning applications. I had envisaged a scripting solution where you simply record your session and when you replay, the system walks you through step by step, pausing for you to do the task with prompts on what to do. Imagine how much easier it will be to train people on new systems. This could forever change the way we train people on new systems. Well, i never did anything about the idea like the other gazillion ideas i keep generating.

Today I was reading the June 2007 issue of IEEE Computer and I came across this article “Social Scripting for the web“:

As our business and personal lives move online, we must learn to carry out increasingly complicated tasks on the Web—for example, checking a bank account balance, setting up automatic bill payments, sharing photos with family, searching for real-estate listings, and ordering new business cards. Wouldn’t it be nice to have an expert watching over your shoulder and showing you how to do such things properly?

The Koala project at IBM’s Almaden Research Center ( is aiming to create the next best thing: a wiki-type repository of instructions for Web applications that can help users automate common tasks.

In addition to being human-readable, Koala scripts are machine-understandable—the system can interpret each instruction and perform it automatically. At each step, Koala shows you what button to push and then does it for you. It can also fill in fields with your name, address, and other personal information.

I am so excited by this and can’t wait for the software to be released to the market. This may be hard to visualize for you. If that is the case, check out this Koala screencast on IBM’s site.

Ganesh, maybe you have access to this software inside the firewall? If you do, maybe you can update us [without violating confidentiality, of course].


  1. Anonymous said June 22, 2007, 1:15 am:

    Sounds good. Hopefully my dad will be able to use the system. He struggles with the simplest tasks, and I end up doing all his presentations for his confrences 🙂

  2. Anonymous said June 22, 2007, 12:39 pm:

    Thanks Archana. I think koala works only on web-based software. So it may not be helpful to your father. Of course, you could try on the web-based presentation software like zoho. -sukumar

  3. Anonymous said June 23, 2007, 12:54 pm:


    I did take a look at an internally available video presentation on Koala. Pretty cool. Without revealing too much, one significant benefit of the tool (other than training/showing people how to perform a task on the web), is the _automation_ part – can interpret each instruction and perform it automatically – For example, navigating through various forms in a site can be now be automated.

    So, besides showing you how to do it, it can also do it. This, I think is one of its USP.

    A nice tool, perhaps not earthshaking, but pretty cool.


  4. Anonymous said June 24, 2007, 5:04 am:

    Thanks Ganesh. I understood the automation part. I also liked the social part – recording the scripts and sharing it with the community.