Barista is Now Fully Conversational!
When we founded Espressive in October of 2016, we had already learned a lot from our experience at ServiceNow about how employees wanted to get help at work. The team had launched the ServiceNow self-service portal with the expectation that employees would gravitate towards such a solution, rather than waiting their turn to speak to a live agent.
But we were wrong. Through this, we learned three painful lessons:
- Most organizations did not have good knowledge discipline, and their service catalog was nearly impossible for mere humans to navigate.
- Employees were generally not technical enough to understand the nature of their problems, and typically could only express their symptoms. Unfortunately, there is no search engine today that can find the right knowledge article for things like “laptop broke” or “VPN not connecting.” The resulting knowledge articles were generally lengthy, requiring employees to self-troubleshoot and repair… which to them felt like homework.
- As a result, employees generally preferred to speak to a service desk agent and have them simply resolve their issue for them.
It was clear to us that to solve this problem, we needed a complete reboot. This was not about making search more effective. Instead, it required us to automate what service desk agents do all day long—help employees resolve issues. Not a trivial task by any means.
Over the last 7 years, we have been successful in delivering a platform that helped our customers automate the resolution of 55 to 67% of their tickets. It was rewarding. Our customers loved it because it enabled them to repurpose their service desk agents to focus on more strategic projects. Service desk agents loved it because it enabled them to uplevel their skills and add AI to their resumes. Employees loved it because they could get their issues resolved in seconds.
And, over those 7 years, we trained our language model to understand well over 4 billion different topics.
In our space, “deflection rate” is the holy grail of KPIs. A “deflection” is a ticket that is fully resolved without human touch. When customers look at our dashboards, they want to understand how many tickets we avoided—this is how they calculated the value we delivered.
Over time, customers also started to appreciate the fact that we were able to reduce their MTTR, which is the amount of time it takes to resolve a ticket. For an employee who is unable to get work done, every single minute counts. So, waiting on the phone for an agent is not something that results in a great MTTR score.
But in November 2022, OpenAI released its latest ChatGPT to the public market, which led to the fastest adopted technology in history (yes, it has since been beat by Meta’s Threads).
And for Espressive, our world changed. Employees who were once happy to engage with our virtual agent to get their issues resolved now wanted to… simply chat. What we started to see during our user acceptance testing phase (where customers validate if the virtual agent is ready for mass roll out) is that interactions shifted away from work related topics to stuff like: “Is a hotdog a sandwich?”, “Are you a bot?”, “What is the airspeed velocity of an unladen swallow?”, “What is the force of gravity?”, “Do you write poems?”, etc.
We started to see a complete change in how people wanted to interact with their virtual agent. Sure, they were not really assessing the value Espressive delivered on our ability to be conversational across any possible topic, but if ChatGPT could be that smart, then their corporate virtual agent better be too. The bar got higher…
And then, all of a sudden, having an understanding of over 4 billion topics meant nothing. This was a wake-up call for sure. And the Espressive team stepped up to the challenge.
I am super excited to announce that Barista has just become fully conversational. Yes, go ahead and ask it anything. Oh, and for those of you out there who really want to push the envelope, be aware we added very specific guardrails to ensure that topics are work appropriate—anything considered offensive will be captured and shared with your leaders. So just keep it PG-13 please.
Exciting times at Espressive. The pace of change is beyond mind boggling, but with the best engineering team in the world, we step up.
Want to test Barista's conversational skills yourself? Request a demo here.
ChatGPT Security Risks - How to Protect Your Enterprise
ChatGPT security risks include sharing of PII, trade secrets, source code, and more - but fortunately, there is a way to protect your enterprise.
The AI Dream Team: Virtual Agent Technology + ChatGPT
ChatGPT is setting records for the fastest-growing user base of all time. So how can enterprises responsibly enable employee access to ChatGPT while at work?
What is Barista Live Generative Answers?
Espressive recently announced Barista Live Generative Answers (LGA), which has raised a number of questions related to how...
My AI Predictions for 2024
This year, my predictions for 2024 focus on the increasing role of large language models (LLMs) in automating issue resolution. In addition, I predict we...
Can LLMs Automate Resolution of Employee Questions? Not on Their Own, But They Can Help.
The level of innovation in the generative AI space has been super exciting, with what appears to be yet another foundational model being announced...