Tagging and automations to scale support reliably
Userzoom is a SaaS company that offers a platform to distribute studies to receive feedback from users (participants) about the products you are selling.
With offices in Palo Alto and Barcelona, I’ve been lucky enough to meet both their US-based and Spain-based team in person before Covid-19 hit.
With Lang.ai, they have seen a significant change in the efficiency of their team and their response time, particularly for urgent issues.
We had the pleasure of interviewing Montse and Narcís, Support Team Managers whose teams help with incidents and requests from both clients and study participants. Their goal is to ensure that their support agents have the tools to operate efficiently to focus their efforts on incident resolution. Additionally, they face the added complexity of covering a broad spectrum of timezones (US, EMEA, and Asia Pacific).
What are the top issues and challenges in Userzoom’s support team?
Our top issue over the last year has been managing a growing amount of tickets as Userzoom is a scaling company, and therefore, issues from customers and panelists keep increasing. One of our goals has been trying to optimize resources and processes so that we can scale.
As a product-led company, the more usage and adoption you have, the better (as it is a daily tool our customers use). But with rising adoption comes an increase in support incidents.
That’s been the struggle in the past year: dealing with this increasing ticket volume while avoiding an exponential growth in hiring new agents. We’ve been asking ourselves the question: what manual tedious tasks can we change?
How does Lang help Userzoom reach your goals?
Lang helps us automate processes as you can have more tickets with the same resources. We’re mainly focused on two areas:
- Automated tagging which means agents are spending less time on each ticket
- Lang.ai’s agent assistant (demo link here) helps us design reminders/notes, reducing errors, and therefore saving time from an agent looking or asking for info on each ticket. When you get a new support ticket, the solution is in front of you, so agents invest less time searching for it.
What was one of the worst experiences in your role before having Lang.ai?
We base a lot of our decisions on the KPIs we define using support data, and therefore they are directly related to the number of tickets from a specific issue. Before, there was ambiguous categorization or agents that were analyzing similar tickets with different criteria. Therefore it was complicated to look at the past and understand the top issues with reliable standards.
With Lang.ai, you’re making decisions with more confidence.
— Narcís Barroso, Product Support Manager @ Userzoom
How has your perception of Lang.ai changed after you started working with us?
Initially, we were very focused on categorization as the primary goal was to achieve auto-categorization. It was once we started using Lang.ai and saw that the accuracy was around 90% that we realized auto-categorization is an enabler for other processes like prioritization or setting up internal notes for agents, which is now super useful.
Who uses the product (Lang.ai) inside your organization?
The Support Management team is the main user, but it provides visibility to the rest of the company.
With accurate support issue classification:
- Our product team can decide with confidence the priorities on what comes next in the product roadmap.
- Account Management/Client Success teams have a view in Zendesk to understand how many times and why their customers have contacted support.
What are the top metrics affected by Lang.ai
- Agent productivity: We reduced the number of interactions per ticket
- Time spent per ticket: Decreases by 1–2 minutes by not doing mundane tasks like tagging
- Ticket handling time/response time: Goes down as agents reduce time searching for information. Besides, as tickets are categorized into different topics automatically, agents can respond in bulk to tickets on a particular topic (payments, for example)
- Our (Support Manager) productivity: We spend less time auditing tickets to review that agents have done an accurate job tagging them. In our specific work, it’s the peace of mind that everyone sees and does what’s required.
- Customer satisfaction: For urgent issues, we have designed an automation with Lang to raise the priority if we detect certain topics.
Even if it’s not always correct, seeing it early makes a difference as we can decide if it needs to be prioritized or not.
We’ve seen that with Lang.ai these urgent issues are being responded 5X faster on average.
I know this quarter has been significant for Userzoom with a big product launch, the pandemic, and the summer holidays. How has Lang helped you during this process?
We already had a remote-friendly environment, so there hasn’t been a big difference in that sense.
However, we had a big release sensitive for many clients, and, as in any release, things were changing daily. Lang.ai internal note automations help us change what the agents see in real-time for tickets about a specific topic, and in this case, we had created a topic for the release (moderated).
This feature enabled us to update the communication daily so that every agent was on the same page. Mostly, as this was during the summer period and some agents were out, it was vital to maintain alignment as we only needed to change a link with the instructions to show agents how to deal with each type of ticket.
What are other platforms in your tech stack for support (apart from Zendesk)? How do they live together with Lang.ai?
For us, it’s Zendesk + Ada + Lang.ai, and the difference between them is clear.
Ada support allows Userzoom to avoid tickets in situations where it’s such a simple question that it can be dealt with by a chatbot.
Lang.ai allows Userzoom to more efficiently manage the incidents that require a ticket while Ada handles the simple queries that can be solved via self-service.
— Montserrat Artigues, Global Product Support Manager @ Userzoom
What three things come to mind when thinking about Lang.ai?
Teamwork, support, and peace of mind.
Thank you Montserrat Artigues and Narcís Barroso for doing the interview and all the innovative workflows you’ve been building with Lang.ai!