In the recent times RPA (Robotic Process Automation) and AI have really caught the attention of organizations. Organizations striving to be more competitive are looking at automation to reduce operational costs while maintaining or even improving customer service. These technologies that were seen as "emerging" not so long ago are rapidly going mainstream. Accelerated interest and investments is beginning to create lucid, user friendly and affordable solutions.
RPA is comparatively a much easier and better understood approach. Automating existing processes which are desktop based, high volume, linear and rules-driven are easier to conceptualize, and possess fewer challenges. Organizations having settled down with RPA are now looking at AI for greater value.
AI on the other hand belongs to the cognitive computing field where it can independently achieve its goal by looking at multiple data points in non-linear manner, perceive the progress and take decisions for alternative paths. Herein we will refer the “narrow AI” as AI to avoid confusion with the higher order, self-aware, sentient AI. It is referred to as “narrow” because its capability is restricted to focused functions like data analysis, image/facial/pattern recognition and self driving cars. IBM’s Watson, Salesforce’s Einstein, MS Oxford and Google DeepMind are some examples of this narrow or weak AI.
This means RPA and AI require different approaches because of their underlying engineering and their areas of value add. While AI can enhance the RPA function significantly, it has much bigger potential.
There are many factors to be considered in order to be successful with these exciting solutions. Key among them would be:
1) Why: Technologies like AI has the potential to create disruptive values for your organization and your customers. We might be limiting this potential by limitations of our vision and plan. Extensive design thinking and planning would be required to get on the right course. Having clear objectives and focused approach would be easier to finance internally and would offer better results.
2) When: While there is no prescribed timing to embark on this journey, simpler projects like RPA can be adapted at any time. For example RPA naturally flows into some organizational functions like finance and accounting. Some might feel that their operations are not mature enough for AI. But to be future-ready, you may want to consider having a small team to experiment with AI and come up with disruptive solutions and approaches.
3) Awareness: Recognize your organization’s technical culture and acumen. These are fairly technology-heavy projects and it is essential to be aware whether your organization has traditionally been early adopter or a cautious player. While RPA can be taken up early in either case, complex projects like AI should be undertaken as per your organization’s technical appetite. For example handicraft industry with little history of high tech would only want to start when it sees better industry-specific AI applications. Since AI derives its fuel from data, organisations might start off by looking at the data and processes and assess readiness.
4) Transparency: Internal resistance to automation is from the perceived threat that the impact of these solutions bring. While blue collar jobs are affected by RPA, the mid to senior management tend to be sceptical of AI even while recognizing its potential. This is not surprizing as AI would significantly change the management decision making process. Bringing transparency in communication and the project objectives would help in managing this internal resistance.
a. System Level: Results of automation solution can be challenged if users and managers do not get insights into how proprietary and often "black-box" solutions are arriving at the end results. Having clear audit trails for traceability of steps and logic used would help build confidence. This would also help in maintaining quality of output and regular tuning of the engine for continuous improvement.
b. Personnel level: Generally AI-like projects have over-arching objectives and are not focused on cutting head count. Avoid fear-uncertainty-doubt by clear communication of the objectives of the projects. Ideally everyone should contribute to the set objectives and should be included as important stakeholders, making them relinquish the status-quo of current operations.
5) Educate and Empower: It would be a mistake to expect our human capital to be auto skilled for not only managing the new environment but be instrumental in creating work processes suitable for such disruptive approaches. For example it requires significantly different skills for a people manager to effectively manage a team with predominant work being done by RPA workers. Similarly managers might find it difficult to be only decision based while relinquishing much of their administrative work to AI. Skill enhancements to address these type of gaps would engage the workforce for better results. At the same time new set of KRAs should be brought in effect to positively shape the performance bell curve.
6) Be aware of the Hype: It is important to be aware of the gap between hype of possibilities and current capabilities of the available technology. For example many so called AI chat-bots are nothing but high level (FAQ) management system with NLP. These bots have no empathy, sense of fairness or justice.
What makes the RPA and AI initiatives all the more interesting is that each organization will have their own use cases and application scenarios. This is creating even wider opportunities for these future applications. Although the hype has superseded current capabilities, both RPA and AI in their narrow bands of operations still have a lot to offer to the businesses of today!
The author is Director Practice, Asia, Verint
Disclaimer: This article is published as part of the IDG Contributor Network. The views expressed in this article are solely those of the contributing authors and not of IDG Media and its editor(s).