For professionals involved in BPS activities, it might be a surprise that software testing (ST) shares a lot of similarities with BPS. Like many forms of BPS, ST is labor intensive – the largest software testing vendors now have up to 30k career testers and 20k has become the new norm. Like BPS, ST has adopted process improvement in a big way. And like BPS, ST is looking at process automation as a way of reducing its labor intensity and eventually move into non-linear growth, improving accuracy and speed as well as efficiency.

It may also come as a surprise to BPS professionals that the software testing industry has already massively invested in task automation, and created a high number of IPs and accelerators. This is such a significant trend in ST that one large IT service firm now says it has ~70,000 testing reusable artifacts.

So perhaps ST has experience to share with BPS? What can we learn?

1. Automation starts small: the holy grail of the BPS industry may be the automation of a whole process but before achieving full process automation, clients can benefit from limited scope and still cut times on repetitive tasks
2. Process improvement is not just about implementing best practices; it goes towards implementing business process modeling i.e. capturing business processes in diagrams. Those diagrams, running on specific modeling applications, are a starting point for automating processes, helping in identifying pockets of manual activities and prioritizing the roll out of IPs and accelerators
3. One area where the ST industry is investing massively is around the dev-to-test-and-roll-out-into-production process known as DevOps. DevOps poses many challenges to the ST community, with its reliance on many different software tools, whether COTS, IPs or open source software. There is another element to this complexity: the traditional alliances with major ISVs are no longer sufficient and the ST industry has had to drive partnerships with much smaller ISVs than in the past, including tech start-ups: something unthinkable five years ago. Some of the innovative vendors have made bets on technology vendors and pre-integrated heterogeneous sets of tools and IPs. The good news is that clients seem ready to accept such pre-integrated DevOps platforms
4. Most automation IPs and accelerators created by the ST industry are free-of-charge to their clients and are meant to drive service differentiation
5. Client organizations do not look overwhelmingly concerned about potential technology lock-up. Credit to ST vendors that have avoided lock-ups and systematically provided integration with major COTS and open sources tools in the market
6. Machine learning, cognitive intelligence (CI) and even AI are also on the roadmap but evidence of current usage is scant. There is some rebranding of agent software sitting on a server or a desktop and collecting data over time to drive dynamically analytics as machine learning.

All in all, we think software testing provides a good number of insights and direction to the BPS industry.

Final observation: has the software industry in spite of its investment in automation become non-linear? Not yet - but this may change with mainstream adoption of DevOps.