There is currently a huge amount of hype around use of automation in BPO, and many commentators seem to forget that the early BPO contracts were often driven by organizations with inadequate/creaking legacy IT systems using BPO as an enabler to combine replacement of legacy systems with an operations and process improvement layer.

These early transformations were also frequently seen by both client and provider as a foundation for utility services, often with the client retaining a stake in the resulting joint venture. So talk of automation and utility services is not new. However, the success rate of these early projects was limited, often being impeded by the challenge, and timescales, of technology implementation and the frequent over-tailoring of initial technology implementations to the client despite the goal of establishing a wider multi-client utility.

So where do we stand with automation today and will it succeed this time around? Let’s focus on the three overriding topics of the moment in the application of automation to BPO: robotics process automation (RPA), analytics, and BPaaS.

These technologies can be assessed in terms of their impact on a business framework frequently used in IT departments (but equally applicable to BPO) of “change the business” vs. “run-the-business”. The area where all the current hype is focused is “change-the business”. At its most extreme this involves organizations in aiming to become more digital and mirror the success of the “disruptors” such as Uber and Amazon. There are a number of technologies that play into this “enabling the organization to become Uber” space, including digital, the Internet-of-Things (IoT), and software apps, supported by analytics.

However, RPA does not belong in this category. Though RPA is one of a number of technologies that are changing the nature and application of BPO, the current implementation of RPA is essentially “labor arbitrage on labor arbitrage” (i.e. moving delivery to even lower cost n-tier delivery locations) in the “run-the-business” category – principally enabling the organization to run with increased efficiency using existing technology and offshoring frameworks.

The final part of this new BPO delivery framework is governance, where Global Business Services (GBS) plays a role in improving end-to-end process management and supporting the integration of outsourced operations with the retained processes or organization.

Analytics plays a role across the piece, principally by supporting identification of process improvements alongside existing lean six sigma methodologies in “run-the-business”, and by facilitating automated process learning in “change-the-business”. It is clear that drill-down dashboards and analytics now need to be built into all BPO services and to increasingly support business actions as well as much more real-time process improvement.

However, current implementations of RPA are typically not cognitive and don’t undertake judgment-based tasks. Though robotics software will often interface directly with core systems via APIs, the current implementations of RPA software largely repeat the keystrokes that a human would perform following a simple set of rules. In essence, that tends to focus RPA around data reading & extraction, data validation and enrichment, and data copying and pasting, along with reporting.

At the same time, RPA largely cannot operate where voice inputs are involved, but is principally involved in manipulating numeric data and structured text.

Accordingly, RPA has so far made most impact in transactional processes such as finance & accounting, and increasingly industry-specific processes in sectors such as banking and insurance, with adoption in insurance ahead of that in banking due to the lesser regulatory concerns. It has had less impact so far in HR processes, though it is applicable to transactional elements here, and its use within customer service is limited, since customer service typically needs more cognitive analytics capability than is readily available in standard RPA.

In the words of one leading financial analyst, RPA can be regarded “as a race to the bottom”, and in an area such as F&A, RPA has the potential to remove a further 20% of the cost base from relatively mature offshored F&A processes over a period of several years. Hence, RPA is having a dramatic impact on existing outsourcing contract commercials. Firstly, the vendors know that their mature BPO contracts in highly transactional areas such as F&A are vulnerable to competitive attack. As a result, suppliers have been rushing to renegotiate their key contracts, and to apply RPA, with typically little desire to remain on FTE-based pricing. If 20% of FTEs are to be replaced with bots, then FTE-based pricing becomes even more unattractive to suppliers than previously. So, suppliers are typically looking to move to fixed price or transactional-based pricing plus built-in productivity targets. The challenge then becomes one of using automation and robotics to drive down the cost curve as rapidly, or more rapidly, than the contractual productivity commitment so as to maintain or increase contract profit margins, with the cost per bot typically quoted as a third of the price of an offshore FTE.

Another commercial alternative is to retain FTE-based pricing within the main contract, but to run each RPA project as a separate gainshare. However, in general rule-based RPA arguably lends itself best to transaction-based pricing, while judgment-based AI technology (when it matures in a few years’ time) will arguably lend itself to more outcome-based pricing.

So, despite all the hype, RPA has a lot in common with the much maligned “lift-and-shift” approach to BPO. It makes no changes to underlying applications and does nothing to improve the overall state of underlying systems. And, as with “lift-and-shift”, it makes operational process knowledge even more important than previously. The key success factors for successful application of RPA are largely identical to those in “lift-and-shift” BPO and include well-defined process and strong standard operating procedures together with agent change management and process improvement expertise. Indeed, RPA is driving demand for lean six sigma consultants, with many of the suppliers employing this type of skill to drive RPA projects to ensure their success. And, as with “lift-and-shift” BPO, one of the key factors in delays is failure to gain the support of in-house IT at an early stage in the project.

However, despite its relative lack of sophistication, RPA has one really major benefit: it makes continuous improvement possible. One of the traditional failings of BPO and the reason why many contracts failed to achieve their promised productivity gains was that the roadmaps to new process models and ability to identify process improvements lacked a quick and easy means of realization. It is one thing to identify process improvements and potential; it is another to realize them if realization requires a significant change in technology or platform implementation, since these have traditionally required significant timescales and capital investment. RPA considerably lowers the time and investment threshold of implementing a stream of modest process improvements with the low cost of a bot license complemented by typical implementation timescales of less than three months.

So, one way of looking at RPA is that it is the new offshore, but one that operates at much lower sub-process FTE thresholds than offshoring and is achieved even more quickly. It can also be applied to retained process elements to improve end-to-end process flow, a common inhibitor to BPO effectiveness, and not just to those processes already outsourced.

In my next blog, I’ll be looking at where RPA, and automation in BPO generally, goes from here.