The impact of AI is largely dependent on the intelligence, analytics and scale that organizations put behind their efforts. Read this article to find out more about IPsoft’s vision for its cognitive AI solution, Amelia.
“We want to bring you return on AI,” said Edwin Van Bommel, IPsoft’s Chief Cognitive Officer, at the second annual Digital Workforce Summit. During his presentation, Edwin discussed the ways in which AI provides lasting business value for companies, and the ways in which analytics, design, and cognitive intelligence helps to differentiate between a potentially impactful AI and a wasted investment. He also introduced a new user interface design and an AI marketplace for IPsoft’s cognitive-based digital colleague Amelia.
“We want to make companies and clients better,” Edwin said. “If you deploy an AI like Amelia, she should help you with that return on AI to justify the investment in the technology.” Amelia is a cognitive AI digital colleague designed to collaborate on business processes for internal and external end users. When an organization hires Amelia (or several Amelias) it will be able to design her to address a variety of company needs, roles, and verticals—all according to specific business processes and procedures..
Unfortunately, Edwin said, many unjustly view AI as a threat, something that will steal jobs or wreak havoc on humanity. AI such as Amelia should be designed to be friendly, he said, and to serve as a colleague rather than as an assistant, in order to help drive true value.
The Determining Factors of AI
“When we talk about return on AI, let’s start with conversational impact,” Edwin said. “You see three determining factors — customer experience, automation, and analytics. Customer experience is the most important. We want people to like Amelia. You want people to come back. This is where you see conversational intelligence and emotional intelligence. We want Amelia to understand you.”
Secondly, he said AI should be connected to a company’s data center in order to provide end users with the most comprehensive set of up-to-date information. In this manner, AI can always provide the highest level of service, in machine speed, with minimal downtime. “Amelia can run an end-to-end process. But to do that she needs to be integrated with your systems,” he said.
“If you’re abroad and for the third time your [mobile] roaming is not working and you connect with Amelia, Amelia sees this, knows it’s the third time, knows you’re a heavy caller and therefore a great customer,” Edwin explained. “Our current analytics tells Amelia to be careful, that this person may leave because he’s not happy, and Amelia can offer you a coupon” or take other measures to rectify the situation.
The more data you have, the more training AI will do, the faster you will get to impact
Volume also determines how impactful an AI implementation can be, according to Edwin. Because Amelia is easily trainable, volume isn’t difficult to accomplish. The more conversations Amelia handles the more intelligent she becomes. As a result, more volume equals more return on investment for AI.
“Volume is something our clients decide. Is Amelia going to handle 50 conversations a day or 50,000?” he said. “Obviously 50,000 gives you significantly more impact. Plus, the AI will learn faster. The more data you have, the more training AI will do, the faster you will get to impact.”
AI and Bi-Directional Feedback
“Why can’t AI steer the full web experience?” Edwin asked. “AI should understand what you’re looking at — a car, a t-shirt? Because then you can have a dialogue about it.”
Edwin demonstrated how Amelia can take feedback from an e-commerce shopper via Amelia and send it to a retailer’s website to parse information (similar to what Google does with a search term). For example: If Mark goes to a retail site and tells Amelia he wants to see blue shirts, Amelia will reload the webpage to show all available blue shirts. If Mark asks her to look at red shirts, she’ll shift the web experience to display red shirts. If Mark asks to see a specific shirt, Amelia will ask for his size and ask if she can add it to his shopping cart for purchase.
An AI Marketplace and User Interface
Edwin also introduced IPsoft’s new AI-focused marketplace where companies can acquire and upload specific Amelia roles. Amelia comes pre-trained to provide services in banking, insurance and healthcare in 16 roles with more than 140 skills. Within these verticals she can accomplish specific customer-service based tasks, such as credit card processing, mortgage processing, and quote-to-cash. Companies can simply contact IPsoft to be set up with a URL where they can download and train Amelia in less than seven minutes.
To set up an Amelia role, companies can use Amelia’s wizard-based configuration process, which is actually led by Amelia herself. Amelia asks for a company name, then she guides users through a UI configuration screen where companies can load their logos, specific color preferences, font choices, among other customizations. Amelia then takes a few seconds to download the intelligence associated with the role she will learn, and she launches. Over time, she will deliver value and business benefits through all of her abilities and skills. Companies can also contact IPsoft to have their own internally-designed Amelia roles added to the marketplace.
“You can start differentiating with Amelia,” Edwin told the DWS audience. “You can come up with your next journeys while she's already deployed and working with your clients.”
More DWS Coverage
- Chetan Dube, IPsoft: The ROI on AI & Amelia Solutions
- Anurag Harsh, IPsoft: The Business of AI
- Max Tegmark, MIT: The Acceleration of AI Is Transforming Business Practices
- Christopher Manning, Stanford University: Inside Amelia’s Brain
- Anthony Abbattista, Deloitte: The True Impact of AI
- Panel Discussion: Cognitive AI in Healthcare Means Augmented Patient and Caregiver Experiences
- Panel Discussion: Lessons Learned from AI Pioneers
- Panel Discussion: How IPsoft’s Clients Innovate and Improve Business with Amelia
- Panel Discussion: AI Deployment Surprises and Strategy