AI Ethics: Building Trust and Transparency With AI Engines

While artificial intelligence (AI) has several benefits for customers, it also has the potential to drive asymmetries further for contact centers. AI engines give you the ability to analyze every call so that the information in the interaction and knowledge can be disseminated to the entire organization.

Real-time transcription, or even post-call transcription, can be combined to better understand the entire interaction. Combining this with the behaviors of all other customers can help predict what might happen next. AI-based engines also can run experiments on messaging and pricing to find a further correlation between behavior and moving prospects and customers toward your specific goals.

New AI tools provide scale and speed that weren’t previously available. Taking this to the extreme, a company can not only change its pricing and messaging, but it can also adjust the language it uses for each customer through technologies like natural language generation. Perhaps the company could even use that customer’s own voice as the text-to-speech synthesis.

In addition, AI engines can inject messaging at the time when it could have the most impact toward a purchase or goal. You can test when a message or action should occur across millions of interactions to find a correlation. Similarly, a contact center that uses real-time transcription can alert an agent of what to say and when to say it so it has the greatest impact.

Agents become AI-augmented humans. It’s like having several many people to consult with on your current poker hand while being able to see your opponents’ cards.

 Preventing Runaway Bias in AI

When applying AI ethics, many companies might wonder if they need to address the imbalance of knowing so much about their customers. When implementing AI, ask these questions to make sure you’re using the technology properly:

  • Does the use of AI provide a substantial benefit to my customers?
  • Can the customer know what we know about them?
  • Can AI engines explain decisions or show reasoning?
  • Can it put the customer on the same side of the table of our company?

It might be difficult for your company to decide these answers. It can be helpful to bring in an independent voice to assign a value to the interaction from the customer’s perspective and then calculate whether the use of AI benefits the customer, compared to what it does for your company. For example, if an insurance company applies AI to determine that a certain customer is less likely to get into an accident and cost them less, then should it sufficiently reduce premiums or just pocket the savings from this information imbalance?

Another aspect is giving customers full transparency and knowledge into what you know about them — and your analysis of interactions. For example, if a company analyzes a call, can the customer be aware that the agent sees this information? If companies are reluctant to share what they do with customer data for fear of legal ramifications or bad PR, it might indicate that they shouldn’t use the data. However, being transparent about the analysis can increase trust — and even benefit the customer because they’re more knowledgeable of their reactions.

To prevent runaway bias and avoid stereotyping, AI-based systems also must provide some explanation into how they reached their decisions and supply evidence to support those recommendations. These insights should also be shareable with the customers they’re meant to influence.

Because the scale and speed of AI allows companies to perform analyses in new ways, you can prevent an imbalance by putting agents and customers on the same side of the table. For example, instead of an agent presenting the results of AI engines as her own thoughts, she might indicate that the system recommends several items. She can then review them together with the client. While this is only a slight shift in how the information is presented, it allows for transparency and prevents undesired manipulation. The customer sees what the agent sees — and the agent becomes a steward of AI technology.

This “same side of the table” approach can increase goodwill and brand awareness for companies. Zero-commission car dealerships, fee-based financial advisors and other professionals who have removed conflicts of interest develop trust with clients. While this isn’t always possible in all verticals, moving agent to more of an advisor role with the AI system creates an honest analysis of a solution or product. Companies can foster more trust and benefit from AI-based applications.

Because AI engines can lead to a huge value creation, it’s imperative to work with your customers to develop the trust that will usher in its adoption. By working transparently and being fair in the distribution of the value AI services create, you can sustainably reap the rewards.

Stay up to date on our AI Ethics blog series and join the discussion on AI Ethics.