If your career focus isn’t technology-related, you might not have had many experiences with Artificial Intelligence (AI). When discussing AI, technology neophytes often reference dystopian examples from science fiction, such as HAL 9000 from 2001: A Space Odyssey, or Skynet and the T-1000 from the Terminator franchise. The chief concern among the under-informed is that AI will one day go rogue and cause massive damage to a company’s finances, branding, reputation, customers etc.
The Better the Training, the Better the Outcomes
IPsoft works incredibly hard to ensure this never happens. Amelia, our industry-leading virtual agent, requires comprehensive preparation before she can go live. Amelia needs to be trained to do the jobs for which she has been tasked — she cannot on her own access data, systems or workflows that she has not been specifically trained to access.
Prior to an implementation, she must be uploaded with an assortment of documents designed to enrich her with new knowledge. A comprehensive glossary of your business vocabulary, employee handbooks, or something as complex as historical customer support data — these are the materials from which our AI develops its basic understanding of how and what we’re training her to accomplish. So, if we upload documents related to financial services, the AI system will become an expert in banking terminology. The same goes for insurance and healthcare. Unless we upload documents related to for example espionage, or cyber theft, or any other nefarious behavior (which we would obviously never do), Amelia will not have even the slightest understanding of how to go rogue.
Amelia: The Process Master
Next, Amelia learns a company’s industry-specific semantic network, and any logic framework she’ll follow in her decision-making. She’ll master a company’s pre-trained classifiers and analytic modules that will help her to conduct business according to industry requirements and regulations. This step of the process is what turns Amelia from a generalist into an industry expert who can maintain company compliance with applicable laws.
She learns to follow the specific steps of a business process and does not deviate from them. In other words: If you teach Amelia to process a mortgage application, she will not spontaneously decide to learn to play the piano. She will approve and reject applications, but she’s only able to do this by processing data and making decisions based on the strict parameters for which she’s been programmed.
Amelia and the Human Element
If Amelia encounters a request that isn’t covered by her training, she escalates the request to a human colleague. When Amelia escalates a conversation, she stays connected with her co-worker to determine how the issue was resolved. With your business’ permission, she will then apply that knowledge to any similar conversations and avoid unnecessarily escalating future issues.
One critical point: Amelia is meant to be used as a complement to human workers. She’s not designed to replace or overrule human decisions. She can automate tasks because she’s informed by data, thus allowing your human workers to focus on more unique issues or complex projects.
How to effectively train AI is one of the many subjects that will be covered at this year’s Digital Workforce Summit on Wednesday, May 8, in New York City. We encourage you to attend and hear how other companies are progressing along their AI journeys. Click below for more information and to register.