Skip to Content

5 Key Concepts Underpinning Effective Conversational UI

Doug Ross
March 08, 2021

Many online experiences now include the use of conversational agents — the artists formerly known as chatbots — that can assist with shopping, banking, and other digital experiences.

You’ll notice, however, that these agents can vary wildly in terms of effectiveness. Some are simple FAQ bots, trained to recognize the most basic of questions:

  • Which is the nearest location to me?
  • What are your hours?
  • Is this part number in stock?

However, questions or dialog that veer from these relatively simple inquiries might lead to a frustrating experience for the end user.

The following are five key concepts that can help drive a more satisfying and holistic customer experience.

5. The Authenticated User: Beyond Simple FAQ Agents

Agents that can securely confirm the identity of the end-user can offer a wide range of additional services:

  • Password reset and account unlock
  • What is my current loan balance?
  • When was my last payment received?

Agents that can authenticate end users, access key customer details and offer transactional capabilities can provide a far more satisfying experience than simple FAQ bots.

4. Sentiment Analysis: Are Users Happy or Frustrated?

AI-based sentiment analysis offers a highly accurate way to gauge end-user satisfaction. An agent that incorporates sentiment analysis can determine whether a user is becoming angry, frustrated or confused.

Sentiment analysis algorithms can analyze the text from a chat or a call center transcript and generate a highly accurate score of customer satisfaction.

Using this sentiment score, the agent can automatically loop in a human agent to prevent the anger from spilling over into an angry social media rant!

3. A Consistent Brand Voice

Brand voice is the “personality” used to interact with customers. Marketing copy, for instance, is carefully written to convey the brand voice. Is your brand voice casual, humorous, sober, professional, etc.? It’s important that conversational agents use that same voice to the extent possible.

This allows for a more consistent — and less confusing — end-user experience.

2. Personalized Agents

One trend we are seeing is the idea of personalized conversational agents. Some financial service firms have created distinct agents based upon the demographics and characteristics of the client.

  1. A Gen-Y investor with assets of $50,000 who trades frequently
  2. An investor nearing retirement who needs help estimating income and when would be the best time to retire
  3. A retired investor with a high net worth who seldom trades and is focused on cash flow from dividend income

These distinct agents hone in on the specific needs of the client with a focus on — ultimately — asset retention, which has a significant impact on revenue and margin for all financial services firms.

1. Human In The Loop

Some conversations simply require the expertise of a human agent. A breed of more sophisticated conversational agents allow automated agents and human experts to share a secure, virtual conference room with the end-user.

This allows simpler and more repeatable requests to be serviced by a bot agent, while specialized needs (say, tax advice for a Schedule C form) to be dealt with by a human subject-matter expert.

In fact, the best conversational agents can help summon multiple human experts in different fields to resolve even the most difficult customer questions.


There is a sixth aspect worth mentioning: continuous learning. As we combine the data gleaned from authenticated chats, sentiment analysis, use of personalized agents, etc., we can assess our overall performance and iteratively improve customer experience.

The systems integrator Capgemini, for example, has developed a methodology and architecture for marshalling multiple channels and dimensions of data to improve conversational agent performance. This framework appears to drive significant improvements in end-user satisfaction as well as help frame innovative new capabilities.

Evolving Landscape

As one might expect, the topology of conversational agents is changing rapidly. Organizations that treat these agents as virtual FTEs requiring investment, care and continuous improvement — rather than one-off technology projects — are poised to drive sustained (and sustainable) business results.

About the author

VP, National Solutions Architect – AI and RPA | USA
Doug Ross is the former CTO at Western & Southern Financial Group, a Fortune 500 diversified financial services company. While there, Ross won a ComputerWorld Premier 100 Award as well as an SMA Innovation in Action Award for innovative solutions that helped the organization open new and highly profitable distribution channels.


    Leave a Reply

    Your email address will not be published. Required fields are marked *

    Slide to submit