Mid of April I published an article about the mobile in-app support landscape that, amongst other players, touched on Zendesk. In this article I stated:
“Zendesk is not a mobile native. Their chat widget integrates into web pages and the company does not offer in-app chat. Instead the company offers solutions that hook into existing messaging apps like Facebook Messenger or Whatsapp.”
This statement was based upon research that I did in the first half of the month with Zendesk publishing their Fabric based in-app support kit on April 19 of the same month. So, maybe I should have posted this article a little later, but good on Zendesk for getting on with mobile in app support. They had, as well as many other bigger vendors in the customer service and call center arena still have, a wide open flank here that gets covered by specialist vendors like Helpshift, Intercom, or LivePerson, or suite vendors like Freshworks.
Zendesk, a Mobile Native or Not?
I say that, although I maybe did them wrong by stating that they don’t do in-app FAQs – although I do not believe so, as the help center content seems to be delivered from the server and needs an online connection.
Still I maintain that they are not a native player. I will explain my reasoning a little later, after summarizing what I got out of talks with Douglas Hanna and, more recently, Greg Dreyfus from Zendesk.
The support SDK enables mainly three things:
- The searching and showing of help center content
- The opening and management of requests (support tickets)
- And the prompting for app reviews
It is what I would call the essentials of customer service, extended to the mobile world.
The chat SDK is what gives the (at least to me) interesting features:
- Attachments to chats and messages
- Chat ratings
- Emailing of chat transcripts
From a customer base, Greg tells me that between 4,000 and 7,000+ customers (4,000 using the Android SDK and 3,000 the iOs SDK) are using the support SDK and a tenth of that the chat SDK. This makes sense as, without further technology, the support SDK helps deflect calls via self-service and keeps requests asynchronous and therefore helps in keeping the service center small. This is in line with Gartner Group’s observation in their 2017 Magic Quadrant on Customer Engagement Centers that most Zendesk deployments are for 20 agents or less, although there are implementations with 100+ agents, too.
On the other hand, the chat SDK is what enables a company to become more customer friendly. It enables the customer to initiate a chat and to leave a message (request, ticket) if no operator is available.
I didn’t get any numbers of active devices that are equipped with either of the SDKs.
What does Zendesk think about AI and Bots?
To Zendesk, AI is an important tool in the shed. However, as per Greg, for the time being it needs to be confined to the back end and predictive analytics, as the intelligence is not reliable enough (yet) to handle customer requests unsupervised. So there are no bots on the immediate horizon, as they would need to be limited to overly narrow areas of knowledge. Having said this, Zendesk is looking into this topic and observing the progress, as also evidenced by their release of their Guide product.
Zendesk is admittedly one of the big kids on the block of customer service centers. Their software is covering all major channels and it offers rich functionality. There are good reasons for it appearing prominently in the G2 Quadrant for Helpdesk Software and in the Leader’s quadrant of Gartner Group’s Magic Quadrant for Customer Engagement Centers. Zendesk has a broad and satisfied customer base mainly in the SMB market. The thriving Zendesk apps marketplace is helping to get even more coverage, or to improve the solution where marketplace partners see opportunities.
So, they are clearly doing a lot of things right and are probably even a good alternative for Salesforce, especially as they have a mobile offering, which is a distinct weakness of Salesforce, which on the other hand has a strong position for customers that do not only look into customer service functionality.
I, however, do think that both, the mobile support SDK as well as the chat SDK can benefit of some improvement. I have the impression that the support SDK pulls help center content from a server. If that impression is true this should be changed to offering the content directly on the device. If I am right here, this is only a minor one, though. After all, cell coverage can be assumed to improve. Still, the search experience is far better in case of the FAQ being delivered directly into the app. Still, Zendesk is working on adding offline search capabilities in order to offer a better experience.
Where I think Zendesk got it wrong is their approach of in-app chat. The way Greg presented the functionality to me it appears that a chat request waits until getting picked up by an agent – which could be any agent, if the customer does not choose to pick one of the offered categories. In case no agent picks up, the customer can convert it into a ticket herself, turning the chat into a message. This seems to come from a strict definition of chat being synchronous and is in my eyes rather service center oriented than customer oriented. It also seems that there is no use of meta information available through the SDK or the text of the request itself to attribute and then route a request into the appropriate queue. This concept appears to me as the extension of web chat into the app and is probably due to the way the back end is designed to work.
Here also comes the advantage of a working bot framework. A bot, basing on a good knowledge base and a machine-learning framework will be able to automate a lot of requests in its domain. It can also support the agents by already providing relevant articles ordered by confidence in an assisting role.
There are vendors that are capable of both. I think that it would be beneficial for Zendesk to become a leader on this way, as these functionalities will become key requirements in the upcoming world of customer service in an ambient computing environment.