It was a great opportunity and experience for me this year to be able to attend the Pegasystems’ annual user conference, namely PegaWorld 2017, held in MGM Grand Hotel in Las Vegas together with more than 4000 people attending from 600 different firms and 50 different countries.

As Alan Trefler, the CeO of Pegasystems, declared in his keynote, PegaWorld was not just an ordinary tech event for me but it was a tech event where I learned about how customers apply Pega’s technology in their day-to-day business.

There were 3 main themes in PegaWorld this year:

1. Agile Enterprise, which is core part of Pega’s strategy and it is not only about delivering rapidly but also about how to deliver and continuous delivery. It is a robust part of Pega’s business.
2. Customer engagement and Artificial Intelligence (AI): It is about advanced decisioning and predictive analytics capabilities integrated into the application. Sales and marketing are based on one platform.
3. Future of Work: Robotics and automation capabilities

In this blog post, I would like to share with you my learnings about AI. My learnings and takeaways here are mostly obtained from the keynote of Dr. Rob Walker, Vice President of Decision Management and Analytics in Pegasystems.

History of AI
Dreams and fears about AI have been there for ages. AI is not a new concept for the world. It has been there since 1940 when Ada Lovelace wrote the first computer program. But why did it take so long for AI to attract so much attraction only today although it exists for years? It is not about the algorithms, the algorithms were already there. It is about the data, the capabilities of using and storing data have only grown recently. Advances in data processing speeds, lower costs, big data volume and the integration of data science into technology has made practical AI a reality for many organizations.

Today, everyone talks about big data. Every day we are storing large sets of data. Data is everywhere and it is too much. This causes data overload which makes the systems slower and makes the work of a customer service agent who is trying to deliver service to the customer much more complex. The problem here is that we are not using the data in the right way. We are treating all the customers pretty much the same. The key is to make this data valuable by using it in the right way. Data will become valuable only if we use it intelligently within every interaction that we have with the customer, if we introduce the right data at the right time and if we achieve to produce proactive insights from the data that we have and if we turn those insights into action.

AI capabilities help us make data valuable by translating data into automated decisions to act, by executing these automated decisions and actions in every customer interaction throughout all channels and by feeding back the AI system back to learn automatically to become better.

Fear of the Unknown
However, AI is not a welcomed concept by everybody. Pega recently did a global survey about AI to find out what consumers really think about AI. They surveyed 6000 consumers in 6 different countries. As a result, it turned out that 1 out of 4 people think AI will take over the world and more of those people think that AI will come after our jobs.

It is not actually absurd to fear of AI when you see and perceive what AI is capable of. We have seen in PegaWorld 2017 lots of examples of robots or things created by AI such as a playlist created by AI, a Rembrandt portrait painted by AI, an empathy robot named Pepper, a humanlike looking robot named Sophia answering emotional questions such as if she is happy with being alive. These examples show the good sides of AI, maybe Sophia not that much. We have also seen how AI mind works while watching Bob Ross, the famous painter, painting in TV. The AI brain makes weird associations while watching Bob Ross in TV. This looks actually horrifying.

T-Switch
This leads us to the consideration that we need to control such AI and make sure that we can trust it. The proposed solution was something called “T-Switch” and T stands for both Trust and Transparency. Thus, we can switch from Opaque AI setting to Transparent AI setting. Opaque AI technologies cannot completely explain themselves to humans. When you use Opaque AI, all weird associations and technologies like neural networks, deep learning, genetic algorithms, etc. all are allowed to be executed in decision making. Although it is very effective, it comes with a high risk. On the other hand, a Transparent AI can successfully explain how it is using data to make a decision or prediction.

Using of Opaque AI or Transparent AI should be a conscious decision by organizations. Opaque AI can be seen as rewarding by some CMO for ex. in marketing, since it boosts the return on marketing budget or harmful by another CMO since it can damage the reputation of the company. Transparent AI can be mandatory for ex. in credit risk since regulators insist on transparency. Therefore, Pega is committed to implementing this T-Switch in all AI software it provides to give the organizations the choice.

We have seen an example of a tweetbot named Tay that was developed by Microsoft using Opaque AI and put on the internet. Tay was polite and nice in the beginning but very soon she got corrupted by the internet and she became a complete racist. That shows us the high risk of using Opaque AI without any supervision which is an unacceptable risk.

Always-On CRM Brain
Pega’ s solution Customer Decision Hub, which is referred as the “always-on” CRM brain using AI capabilities to work with the data collected from all channels inbound and outbound to create intelligent, personalized customer engagement, provides a Model Transparency feature which can be set from 1 (very opaque) to 5 (very transparent) to give the organizations the choice to be able to adjust the transparency of AI used. Pega operationalizes AI technologies through the Pega Customer Decision Hub and they then service it up inside Marketing, Sales, Service. In each of these areas, different AI technologies take part in different levels. For example; decision management is equally important for all of Pega’s CRM applications. In Sales, deep learning is less important than Marketing because we need the why. In Customer Service, Robotic Process Automation and Virtual Agents become very important. The T-Switch is also available for all of these CRM applications.

Conclusion
We see that AI has huge potential to augment the human judgement. Therefore, instead of competing with AI or using not welcoming words, we need to try to be more cooperative and boost that potential of AI. With the control mechanism that is mentioned above, we will not only drive unprecedented outcomes but the outcomes will also be safe, says Dr. Rob Walker. Therefore, we need to use AI as a supplementary and complimentary solution to enhance and support human productivity throughout all aspects of personal and professional life.

References

1. AI in Customer Engagement: Balancing Risk and Reward, Dr. Rob Walker, Vice President, Decision Management and Analytics, Pegasystems
2. Ever Wonder What’s Inside that “Always-On” Customer Brain – The Pega Customer Decision Hub, Maarten Keijzer, Product Management, Pegasystems & Ionut Rusu, Product Management, Pegasystems

About Selda Canbaz

 Selda BPMConsultant BPMCompany

 

 

 

I am an Integration and Business Process Management (BPM) Consultant with experience in Integration, BPM, Workflow Management and Case Management systems. I have more than 7 years of experience in IT. Currently, I work as a Pega Developer in client projects and recently I got interested in using AI technologies in decisioning and data management. I am passionate about using BPM and decision management capabilities in order to drive business value by optimizing and digitizing IT operations.