IBM Think: How Watson Works Inside the Platforms You Use Today

Embedding Watson in a software platform gives you a chance to create a “great marriage.”

Four companies are doing this well:

• DXC and Customer Care
• SAP Ariba for Procurement
• LivePerson with RBS for Customer Care
• Salesforce and Watson with Orange Bank for Customer Care

DXC and Customer Care

DXC isa combination of CSC and HP Enterprise Services. We want to answer our customers questions. We like to trust the IBM platform and use a bot in customer care. Most of our customers are Fortune 500. We want to make customer care more efficient.

What’s next? We want to take the bots and tie them together. Perhaps a bot of bots.

How do you get this solution? A lot of this involves integration. DXC does the integration as we implement the solution for you.

SAP Ariba

We automate procurement in the enterprise. The transaction value in their network is more than Amazon and Alibaba combined. SAP Ariba wants to bring the entire context of the user to bring about actionable insights. We want to take internal and external data and bring it to a better value.

We are looking to embed IBM Watson APIs in the products we are building. For example:

• Contracts: many are in files and other unstructured formats. It’s hard to parse through the contracts and get to the right information.  Using IBM Compare and comply API to figure out which contracts you can use to do better compliance, terminating contracts that meet the right criteria etc.
• Sourcing: How do you bring in external intelligence to give them information in their categories.  For example, the cost of a raw material decreases.
• Supplier Risk: Expanding their current product to understand the impact of a natural disaster event.  Understand how to mitigate the risk.

SAP Ariba wants to bring better business outcomes, retain knowledge, and unlock the data.

LivePerson with Royal Bank of Scotland

Royal Bank of Scotland has 16 million customers and has been around for 300 years. We have relationships across generations. Historically, it was face to face and personalized with personal relationships. Over the recent past, we evolved to the digital world and lost that personal touch. Now we can use a digital system to be the trusted banker for our customers.

RBS uses an AI chatbot. The virtual agent allows us to fail fast. Each day we look at what worked and what didn’t and how we can improve it. We can implement those changes immediately.

We started the journey into AI by talking to customers.  We asked how can we help, how can we engage. In the past, if someone had a dispute, they wrote a letter. Now they send a tweet.  our Live Person platform engages in the right channel. You don’t have to reach out one at a time.

The solution unpacked. We leveraged the LivePerson solution and added “Cora” into the system. We started with routing 10% of traffic to Cora and continue to grow the traffic.  We measure how well Cora does each day and continually work to improve the AI.

What’s next? RBS is excited by the future. We intend to expand to voice interactions. We just added an Alexa skill where Alexa is just a channel. We are at a 50% rate where no human interaction is needed. We want to continue to grow that.

We also want to start to use Watson for an internal help desk. The name for this AI is Archie.

Salesforce and Watson with Orange Bank

This partnership has been in place for a while. It enables customers to make smarter and faster decisions. Watson knows structured and unstructured data throughout the enterprise. Salesforce Einstein focuses on predictive analytics in the structure of their cloud.

Orange Bank uses Watson AI to power interactions with 100,000 customers. It’s called Djingo and you can interact with it via a variety of channels. While Salesforce is a core system, they chose Watson for the customer care option. Today, about 50% of customer care interactions go through Djingo. They want to grow it to close to 100%.

If the customer interaction isn’t working, Djingo transfer the session to a customer service rep. The transfer is to Salesforce and the live agent. this integration gives the view of the data in salesforce, what the client sees in the mobile application, and the chat history. This makes the transition seamlessly.

We are increasing the number of ‘intents” and giving more personalized transactions.

The next step is to let customers perform transactions. “Do you want me to block that card for you?” and then it’s done. They will do a number of transactions like:

• increasing credit limits
• transferring money

The speed at which Djingo is learning is incredible.

What’s next? At this stage, we have the key products and functionality. We want to increase the wow effect. It will be a combination of technology and agent training.  It’s technology and people.