How TeamSystem Reduced Workload by 60% and Boosted CSAT by 20%

[Chapter 1: The metrics behind their success]

[Music] we all know support chatbots are but your results with customerly AA are incredible carola how did Aura completely transform your customer support sure since integrating Aura our resolution rate has gone up to 60% our average waiting time has been reduced to 12 seconds which is a big improvement from before which was 2 hours we handle about 2,000 ticket daily and we order the workload on our human team has decreased significantly as for cisat our customer satisfaction scores uh Aura has improved it by 20% this is actually higher than when only human agents were handling queries wow a 20% boosting customer satisfaction is no small fit why do you think your sear improve so

[Chapter 2: Why CSAT has improved +20% by using Aura?]

much after implementing Aura I believe it's because Aura can respond faster and more consistently than humans especially for repetitive questions customer appreciate the immediate responses and the accuracy for the information they receive also aa's ability to escalate complex cases to the right human agents ensure that no time is wasted and the right expertise is brought in at the right moment there are a lot of AI tools out there but not all of them deliver results what made customerly Ora stand out out from the competition for your

[Chapter 3: Why Aura was chosen over other AI solutions]

team customerly Ora stood out for several reasons first is of importing knowledge Ora made it seamless to integrate our existing knowledge base which saved us a lot of time during setup second flexibility mapping user intents or goes beyond first level knowledge we can design complex missions that automatically direct customers to the right Solutions even for more advanced inqueries third automation of back office processes Ora can gather important information and pass it through apis helping us automate tasks uh that use to be manual keeping an AI assistant up to dat can be tricky especially as your product evolves how do you make sure Aura stays current with the latest updates to your knowledge

[Chapter 4: How do you update Aura?]

base our content team stays ahead of updates at furing Cloud even before a new feature is officially released we already have guides and can respond is ready once the release happens we publish the content and immediately train h on it this proactive approach ensure AA is always up to dat answering um customer queries correctly when a

[Chapter 5: Update Kb VS Creation of Intents]

customer questions comes in how do you decide whether it's better to update an existing knowledge based article or create a new intent in AA we follow a two level approach first level if a query can be solved with existing guides we update the relevant articles in the knowledge base second level if it's a more complex query that requires specific actions like sending the customer to a team or automating a process we can create a new intent or Mission within AA we all know that chat can sometimes struggle with similar or ambiguous queries how do you train AA to

[Chapter 6: How do you train Aura to give accurate responses?]

differentiate and deliver the most accurate response for instance if a customer asks about aound or needs to figure out if it's about building aound or something else by reviewing how Aura handles similar queries both right and wrong we fine tune its decision making so it get better over time in the early days of using Aura how did you figure out when it needed more human intervention to improve its performance

[Chapter 7: When human intervention is needed to improve Aura?]

at first we had to people constantly monitoring aura's responses whenever AA made a mistake we would review the issue and either update the knowledge base or create a new intent or Mission now we don't monitor every single response but focus on spot checks and feedback through cisat if we notice a trend in negative feedbacks uh we retrain AA when

[Chapter 8: Which signals you monitor to retrain Aura?]

you're monitoring aa's performance what are the key signals that tell you it's time to retrain it or refine it our uh check Che AI group plays a key role when a query ends up with a human agent we analyze why if we see that the same questions keep requiring human help that's our signal to retrain Aura whether it's updating the knowledge base or creating new intents or Mission what were the biggest challenges when setting up Aura and how did you solve them one of the toughest challenges was knowing when they update an article versus creating a new intent or Mission at first first I struggled to make those decision and later the team did too over time we developed a process we tried to updating the knowledge base first if fora still didn't response correctly we create an intent or a mission this trial and error approach combined with

testing helped us to get faster and solving these issues customers love fast automated responses but they also value

[Chapter 9: How to balance automation and personal interactions?]

the human touch how has H helped you balance automation with c some expectations for personal interactions with Hora automating the first response to every query we've saved a lot of time for our team now we are even exploring the possibility of having phone support some things that we never had time for before this way we maintain the human touch where it's most valuable without sacrificing efficiency how does Aura determine when it's time to escalate a

[Chapter 10: When Aura escalate?]

query to a human agent and what does the process look like cor escalates when first whenever she doesn't know the answer second a customer explicitly asks to speak with a person third an intent or a mission is triggered that requires human intervention or escalation to a specific team MX tell the real story of success so what key performance

[Chapter 11: Metrics to monitor Aura Performance]

indicators do you track to ensure Aura is performing at its best we monitor three main metrics first customer satisfaction We compare oras cisat with that of our human agents second the average response time we track how quickly a response versus human agents third human intervention rate we analyze how often uh our query gets escalated and adjust hor training to reduce this cool so looking to the Future what Advanced AI features are you planning to

[Chapter 12: Implementing Advanced AI]

integrate with aura to handle even more complex workflows we are planning to integrate more API automation to handle back office tasks uh these tasks don't add much value when done manually so automating them will free up our agents on focus on more complex issues any example maybe you want to share first automating the found process from uh zero to closure which means that um we collect all the data from the customer in a conversational way with the mission then we connect customer Leora to our API to check for eligibility if everything is in order we issu the refund and close the conversation directly which is wow before AA a refund ticket could have been taking has uh one week to close another incredible Advanced flow we manage with API is the increase of invoices allowance on customers plan we connected a to our database to increase the inv allowances and we close the ticket in less than a minute [Music]