On the 13th of June, giosg hosted a webinar "Improving Live Chat and Chatbots with AI", focusing on the application of artificial intelligence (AI) in customer service.
In this summary you can find:
The webinar host Kimi Tiinus, giosg’s Product Marketing Manager, started the webinar off with a poll to gauge the audience's current adoption of AI in their live chat or chatbot systems.
The results showed that only 15% of the people attending the webinar are already using AI, whereas 55% are not utilizing AI. But altogether 30% have plans to implement it in the future.
The webinar began by highlighting the importance of combining automation with meaningful interactions to provide a seamless customer experience.
As a starting point, organizations have the (optimal) ultimate goal to solve 100% of customer queries.
To reach this, Tiinus introduced button bots as a solution for automating responses to simple questions, covering around 50% of customer queries. But to be able to automate up to 80% of customer queries and be more capable of handling complex questions and engaging in conversational interactions the solution lies in AI solutions and specifically chatbots as the next step for development.
The main message was that to achieve the full service potential of 100% there needs to be a combination of AI chatbots with human agents, aka a live chat alternative.
Improving live chats and chatbots with AI: Case giosg & generative AI
In relation to this, the webinar also introduced a case study of a Finnish rental apartment company that utilizes giosg’s recently launched AI solutions to increase its service efficiency.
The company achieved a 50% automation rate by utilizing giosg’s chatbot solution.
Tiinus then went on to share another use case to further improve automation and service efficiency. Instead of relying solely on a regular knowledge base, companies can utilize generative AI solutions, like giosg's, that utilize chatGPT and can be used to for example summarise long answers from articles into shorter responses.
This allows for more customized and human-like answers.
This type of solution brings numerous benefits, such as less manual work, maintaining a natural tone of voice, providing more specific responses, and delivering efficient in-context replies.
Throughout the exploration of generative AI in customer service, Tiinus introduced giosg's recently launched Live Chat tool Copilot, which provides live chat response suggestions to the human agents that allow faster responses that fit the matter in hand without having to spend time manually formulating each response.
This feature also uses generative AI to offer response options based on the visitor's question. CoPilot offers improved language support, reduced manual work for agents, and ensured more in-context and natural-sounding conversations.
In conclusion, the webinar touched on the future of AI in customer service, predicting the emergence of more diverse AI solutions and better tools for agents, while emphasizing the numerous benefits of incorporating AI in customer service:
✅ Better first-line support and self-service through chatbots
✅ Improved resolution rates
✅ Personalized and customized support
✅ Faster human support through live chat generative AI (CoPilot)
✅ Customized customer experience.
Going forward, five years into the future, here's what we're looking for from AI in customer service:
🤖 More diverse AIs
🤖 Better AI tools for agents
🤖 Enabling automation through entire organizations
Overall, the webinar provided valuable insights into how AI can enhance live chats and chatbots in customer service, showcasing real-world examples and discussing the potential for future advancements in the field.
1. Is it easy to implement and "teach"?
Yes. The AI solution can automatically import and use different kinds of content, while the manual entry of content and review & additional QA by humans is still also possible.
2. Is it the same quality of responses also in other languages than English?
The generative AI models are currently mostly optimized for use in English, but many of them are robust enough to generate good results for a variety of languages.
3. How is it possible for the AI to use CRM, e-mails and even phone call data in terms of GDPR?
This depends on the use case and the actual data. When AI Copilot is used, AI is not directly communicating with end users so there is no risk of it disclosing any personal data to which GDPR applies. Content could also be reviewed or automatically scrubbed beforehand to remove personally identifiable or otherwise sensitive information.
4. ChatGPT and GDPR? Where is data stored?
5. Is it possible to teach GPT the tone of voice of the organization? Or does the tone come from a source that GPT uses?
The AI can be verbally instructed to use any particular kind of tone of voice.
6. Is it possible to define which website the generative AI uses as a database to give correct answers?
Yes, it is.
7. What would be the average cost to take AI Chatbot in use: licenses and the tool etc?
Giosg's AI solution pricing consists of a one-time set-up fee and a monthly recurring fee. The fees depend on the particular AI solution chosen, the volume of chats and (custom) data sources used. Please contact giosg sales for more information.
So if you'd like to discuss this in more depth with one of our experts, get in touch with us!