Web Summit One Year Later: What We’ve Learned

By Clustaar October 31, 2018
Websummit what we have learned Clustaar

Clustaar, a customer support automation platform will be exhibiting at the Web Summit for the third consecutive year. The platform launched last March and now has 1,800 accounts from nearly 85 countries.

In the past year, we have invested resources in selling the product and in customer success organization. Many of our clients are anchor clients that have shared key insights with us which allowed us to get key chatbot and automation insights.

As we prepare for Web Summit 2018, the largest startup & tech convention in Europe, we would like to take this opportunity to share what we have learned during this past year, and how we see the future of customer support automation.

Customer Support Needs Automation

Through speaking with potential clients, startups, and investors, at conferences along the way, we have narrowed down our core business use-case. Our conclusion is this: There are two types of convincing use cases for bots. One is where the bot reduces existing costs (or avoids extra costs)  and the second is where the bot generates additional revenue (or avoids lost revenue).

We believe that the most convincing use case is where the bot reduces existing costs. Reducing cost is more measurable than pinpointing the factors of increased revenue.

Most of the companies we talked to had the same CS issue: their customer support teams cannot keep up with the increasing number of user requests, and most of those questions are all the same. For software companies are all looking for scale and human customer support doesn’t scale very well. By automating all recurring questions, it relieves customer support teams, which allows for more time to tackle real issues that demand human intervention.

For a password-reset request, or for a simple question of “where can I find this,” a bot can be very effective.

Imagine a bot that manages to save 500,000€ in a customer support case. It’s 500,000€ that goes directly to the bottom line. Whereas it will always be difficult to attribute 100% of a 500,000€ increase in revenue to a bot. Moreover, 500,000€ in new revenue will only contribute 15% of that to the bottom line.  

B2C Customer Support Is Best

When it comes to customer support automation, the more users with support queries the better. Especially if the queries are simple level 1 request that can be easily answered (i.e- I need to reset my password). Usually, in these type of situations, you get a Pareto effect: 20% of requests will generate 80% of volume.

Most Companies Do Not Know How To Build Bots

Over the past year, we have discovered that providing companies with the tools to build bots is not necessarily sufficient. Often times companies lack the right experience and build bots that do not deliver sufficient value to the end user. For example, bots that provide the latest news or movie schedules, things that users can easily complete by themselves with a web page. This shows that bot-building skills are in short supply. In other words, not only do companies need bot technology but they also need training in using it.

When we built the Clustaar Platform we had one objective in mind: make it user-friendly and easy for clients to use without assistance. However, reality check: most companies would like help in the process. So we built a Customer Success team to follow up with clients, train them on the Platform, help them to draw solid scenarios, and follow the project to improve the bot along the way.

Today, we sell more higher price licenses which come with support and assistance than licenses without help.

Another tip: given the still-low understanding of bot technology in the market, it is better to choose the value-oriented projects than innovation-led projects. When we meet prospective clients, these are the first 3 questions we always ask:

What is the value chain that you want to improve?

What makes you think a chatbot is a viable solution?

What is the key KPI by which the bot’s performance will be measured?

If the client cannot provide a clear answer to these questions, then the chatbot project is most likely half-baked and will lack support and funding internally. It’s wise to run away from that.

Live Handoff is not necessarily best

There are situations where clients would like the conversation to be escalated “live” to a human agent when necessary. In these cases, the bot takes the request to a human agent who can respond right away.

A live handoff looks deal, however for businesses, this is not always possible. The live handoff involves time and resources which may not be available. We have also found that the live handoff is not always necessary. Users are content with the promise that a representative will reach out to them at a later time.

For live escalation to work companies need the right resources, i.e – a call center sized support team able to take the additional questions that the bots are sending on the spot.

For example, if you are a fast growing startup with a team of 4 customer support agents, it may be difficult for those agents to cover all requests, around the clock. In this case, a live handoff may not necessary nor recommended. Daily, requests come in surges and it can be difficult to process them with consistency.

Now imagine the same situation where all end-user queries to go through a support chatbot, which answers 50-60% of requests and the remainder is transferred to the support team. When these agents receive the bot-prequalified requests, they now have the time to properly process the queries. With the advantage of having the queries prequalified by the bot, the team can provide complete and appropriate answers much faster than before. The result is overall better customer satisfaction. Take a look at our Web Summit 2018 takeaways in our blog post

Over the past year, we have gained valuable knowledge of how to provide user-friendly and efficient bots to companies and their end-users. The platform makes it easy to set-up and maintain bots. Try it for yourself, or you can just request a personal demo from our team! 

Build your own chatbot