21 Sep Conversational AI For Customers and Agents
- How a Conversational AI interaction works
- How CAI can Help You and Your CSAs
- Key KPIs to Measure
- The Importance of Transparency, Sentiment and AI Evolution
Connecting with your customers on a personal level is important to building lasting relationships. But individuals only really hold one focused conversation at a time. Conversational AI (CAI) helps with the scaling of personalized customer attention and interaction sought across industries. CAI is a growing trend with a worldwide market projected to grow substantially from USD 6.8 billion to USD 18.4 billion by 2026.
How a Conversational AI Interaction Works
CAI applications include chatbots, virtual interactions and voice bots. This technology relies on machine learning and deep learning to parse queries and apply appropriate responses/solutions.
The basic path for the AI in a customer interaction:
- Customer enters query
- AI breaks down the query for meaning
- AI provides an appropriate response
- If the AI can’t decipher the meaning or form an appropriate response it escalates the query to an agent*
- When feature is available – AI provides QA for the interaction, including measuring sentiment
* Additionally, CAI can be used to assist agents in real time during customer interactions
To achieve this, multiple layers of analysis occur including natural language processing (NLP) and its subprocesses, natural language understanding (NLU) and natural language generation (NLG).
The effectiveness of this overall process is dependent upon the robustness and relevance of the datasets provided to the AI and troubleshooting to create a smooth process. Just like quality agent training is vital to agent performance, so is quality AI training vital for AI performance. The importance of investment in training and troubleshooting is key to preventing the damage of a bad first impression.
Note, some chatbots are still rule-based programs. Zendesk aptly describes them as similar to an automated phone menu. For certain simple tasks they may suit your needs.
By contrast, a CAI chatbot continually learns and evolves to offer better answers over time. They are a long-term investment toward personalized automation.
How CAI Helps You and Your CSAs
To start, conversational AI can respond to simple inquiries that your business gets most often from customers. This allows your agents to focus on more nuanced issues. It also reduces your staffing needs as your bots can be available to your customers 24/7.
CAI can also act as a filter. If it can’t find an appropriate response to the customer query after applying NLU and scanning your customer database, it can escalate the query to agents in queue.
Further, there are advantages to internal facing CAI directly supporting CSAs, like enabling agents to query the AI during interactions. This saves time, prevents having to put customers on hold or tying up a manager.
Key KPIs to Measure
To measure the effectiveness of your CAI approach and cost savings consider measuring these KPIs:
Self Serve Rate (SSR) – Customers that interacted with AI only, not CSAs. Use this to compare against your CSA interaction rate and overall service rate (inclusive of both) to help measure customer adoption and service effectiveness.
Average Cost – In this case, taking the average cost for an agent assisted interaction multiplied by the number of CAI only interactions for measuring the cost savings of your CAI interactions.
Consider comparing Time To Resolution (TTR), Average Resolution Time (ART), and Average Handle Time (AHT) of your CAI interactions vs your CSA interactions to measure effectiveness.
The Importance of Transparency, Sentiment and AI Evolution
The aim is detailed, accurate interactions with a human touch. However, the more sophisticated the AI the harder it is for customers to tell whether they are interacting with a bot or a person, at least initially. A seamless exchange is the goal, but not at the expense of transparency. Customers want to know if they are talking with a bot, not discover it partway through an exchange when the bot fails to respond in a ‘human’ way. That can cause an otherwise positive touch to sour.
This is one reason consumers have mixed feelings towards AI. The accelerated adoption of AI tech in 2021 revealed many of its ‘growing pains’. Bias, lack of empathy and a disconnect from human experience are not what you’re aiming for in your customer service strategy.
Data is showing a changed but nuanced reaction to bots and AI. In 2019 Forrester released a report indicating that consumers don’t trust AI, but that it has a place when used wisely. However, in 2021 Forrester released another report indicating AI as one the fast growing megatrends. AI magazine predicted a further increase in AI use after the rise in development of Responsible AI, which includes Explainable AI, Ethical AI, Auditable AI and Humble AI, developments that improve AI decision making. Humble AI is purposefully not certain of the right answer, limiting the AI’s confidence and scoring it.
Pullak Mohanty, founder of Navedas, makers of CSAT.AI says, “Successful CX outcomes utilize sentiment analysis to augment current conversational AI. Customers feel that both their needs are met and voices are heard. Brand voice phrase-based models and CSAT.AI’s empathy feature enable AI to read your customers’ sentiment, advise agents accordingly and simultaneously provide QA in real time. For more complex customer service issues CSAT.AI assists live agents in real time.”
Many customers understand that technology is helping them to accomplish tasks, including acquiring customer support. AI based service is an increasing trend companies need to adopt to remain competitive, to scale and to reduce long term costs.
Plan your company AI adoption carefully to meet customer needs with speed and intelligent personalization.