How Automated Translations can Power a Multilingual Jira Service Management
Is excellent support at a global scale even possible?
The standard of excellent customer service keeps moving upwards. You certainly need an agile, ITSM compliant platform like Jira Service Management (formerly known as Jira Service Desk), but there’s many other factors. Problems must be solved within minutes without going back and forth between agents. Customers will avoid throttled hotlines at all costs, but at the same time they want to be able to reach a qualified and empowered agent on the phone when the matter gets serious.
Customer experience must be consistent independent of the channel. The growth of chat communication, self-service portals and social media interactions have made the 8 to 5 schedule obsolete. You will need to be anywhere, any time. Whistle and I’ll be there for you 24/7.
The number of companies taking an omnichannel approach has increased six times in the past five years
Zendesk
Customers want to be in control. This is particularly true of critical applications like Fintech B2C products, but it really extends across the board. It’s a difficult game, and your reputation is at stake.
Tearing the language barrier (externally and internally)
47% of companies identify sourcing and agent retention as primary pain points of their multilingual customer support
In the current global context, forcing users to communicate with your brand in a language that is not their own will alienate them. Brands that don’t localize their support can be perceived as unfriendly. Unfortunately, creating local support for every country where you have customers is impossible: there’s simply not enough talent out there, and the margin of your products can only pay a maximum number of salaries.
Language is not less of a barrier internally. Instead of collaborating based on their domains, your experts work within their linguistic comfort zone. No matter how much you train them. Particularly in some nationalities, they will even avoid to work in tickets in English. Over time, you support backlog can start growing out of control. How can you make those agents work on support cases they would otherwise have no access to?
Machine-learning multilingual translators like google translate, Babel Fish, Unbabel or Deepl provide real-time translations and can play an important role in breaking down those linguistic barriers. If you also want to deliver a truly global support experience on Jira Service Desk and enable your team to collaborate across boundaries, you should consider enabling some sort of automated translations. But be careful not to rely blindly on them, as they can’t substitute the voice, tone, and accuracy of a human agent.
Organizations providing strategic multilingual support will gain important market share for providing interactions that are easy and solve the customer’s problem, regardless of language.
Unbabel
There are some early adopters of these kinds of hybrid strategies in the Atlassian world, but the projection is that in 5 to 7 years all customer interactions will be assisted by AI to remove language obstacles. For example, Tickets Translations for Jira Service Desk can embed automated translations in your Jira Service Desk tickets and communications. Powered by Google Translate, you will have full support for 109 languages in all their combinations. Multilingual customer support can be a great competitive advantage and an incredible customer satisfaction booster.
(for full disclosure, Ticket Translations is developed by resolution, a Gold Marketplace Partner and my employer).
How are you going to use all that power?
After conducting many customer interviews, we have identified 5 key scenarios where automated translations will bring an enormous value to your support team on Jira Service Management.
Scenario 1: Provide level 2 and level 3 support in every language your customers speak
In an ideal world, your international customer service team will have level 2 and level 3 support agents that speak every possible language. In that situation, nobody needs to translate a customer email or a ticket. The real world is a bit different. It’s hard to find experts that have both the linguistic and the technical skills to be level 3 agents. And even if you could afford to hire them and managed to retain them over the years, in low level languages you might only have three or four tickets per week on their Jira Service Desk queue.
Instead, you can hire qualified agents that speak the strongest language and rely on a multilingual translator. The automated translations will break down the language barriers across levels of support and ultimately create a quick qualified communication with the customer. Now you’re ready for tickets to flood in.
Let’s look at an example of communication flow.
1. Incoming customer request
- Camila, a Portuguese Level 1 agent needs to escalate a ticket, but there’s no Portuguese Level 3 agent.
- The ticket is elevated to Kate, an available English Level 3 agent. Camila adds some comments for context.
2. Ticket resolution
- Kate then uses Jira Service Management and the embedded ticket translations to read the customer’s message and Camila’s notes
- Carmen finds a solution and sends it in Spanish to Camila, who uses the translation as a starting point to draft a Portuguese message for the customer
Scenario 2: Eliminate waiting times for internal Help Desks
Employees who don’t feel comfortable using a second language can go to great lengths to avoid embarrassment.
To better empathize with the situation, let’s highlight it it with an example.
Imagine Toshiro, an accountant in Kyoto who works at the national branch of an insurance company headquartered in London. Toshiro interacts with his colleagues in the accounting team and with local colleagues from other departments. Although English is the corporation’s business language, the local work is entirely in Japanese. In fact, Toshiro never has to speak English, which is not even a job requirement for his position.
Whenever he has an issue that would usually mean he needs to communicate with the company’s IT customer support, he instead, fixes the problem himself or asks his colleagues. He’d rather not communicate in English with the central help desk.
Toshiro’s laptop has been slowing down, but his spreadsheets still open. Toshiro would have asked long ago about the laptop replacement policies. He doesn’t want to demand a new device: he wants to know what he can expect, and he knows how to make this ask in Japanese. IT regularly gets these requests through Jira Service Desk. They check device models and manufacturing years (Toshiro’s is a proud 2009 specimen), meaning that the replacement of his dated laptop is approved in no time.
Note that the barrier is linguistic and cultural, but translations can help alleviate the tension.
With automated ticket translations, employees can communicate with IT in their own language. They don’t need to know how to translate a support email into English – they’ll do it with a button. Simple requests like password recovery and device replacement will be solved in the first touch, while complex problems may have to be escalated to involve bilingual colleagues. But all in all, this process will surface many unknown issues, increase local productivity, boost employee confidence, and give management a better IT diagnosis.
Scenario 3: Ensure multilingual support in non-peak times
If your company acts at a global scale, you will need a plan for 24/7 multilingual support . Although customers experiencing problems on weekends, at night, or on a different timezone demand the same support quality, low volume tickets and night-shift fees will only make a small team possible.
Imagine that you could drop the language factor.
How many agents would you need to work those JSD queues and delight customers? Let’s assume you usually cover 8 languages for all 3 levels of support, with a minimum team size of 24. You may want to have at least one level-3 guy for each of your 3 main languages. Supplement with a diverse background for levels 1 and 2, and you can downsize to as little as 8 agents. Since level 1 support will cover every language, customer communication will preserve a human touch.
Automated multilingual translations can speed up team communication and break down collaboration barriers. A smaller team will find the solution to your customer’s problems much faster than with bilingualism or human translations.
Scenario 4: Expansion into new markets with remote support
Are you in an intangible industry like software or services? Then you may expand into new regions and countries without opening local offices or even having direct contact to the customer. Perhaps you didn’t even write a go-to-market document. Customers simply started buying your product in countries whose capital city you can’t even name. And it’s a similar picture when companies in more traditional industries want to expand to new markets. First, open a small office to test the waters. Second, get customers. Only then will you start thinking of providing local support.
Guess what? Those new local customers might not be your prime target, but English support turns them off. If it’s premature to hire a native speaker to keep them happy, you should still provide multi language support via email with automated translations. While your volumes are low, you can bootstrap this. Have an account executive or local employee proofread the outgoing messages, and customers won’t get robotic responses.
Scenario 5: Create easy insights for managers
Running a global support workforce is no easy task. To ensure consistency and quality, a central management team is usually in place to check the content of the requests and the flow of communication between customers and the Jira helpdesk. This central management team will keep many metrics under control (including Jira Service Desk’s native SLA reports), but will keep a special eye on First Contact Resolutions (FCR).
FCRs are the new golden metric for measuring customer satisfaction, so understandably every support manager wants to move the needle. It’s tough. To increment first contact resolutions you need to dig deep into the Jira Service Desk tickets, analyze them, and find common patterns that explain why some requests are not solved immediately. That’s when the underlying processes can be framed and eventually improved.
Managers need to have full visibility, but they can’t speak every language. Additionally, keyword analysis and keyword-based reporting can’t handle content in multiple languages. This makes it even more important to read the tickets created in long-tail languages, where the support team might be newer or less specialised. In other words: quality assurance is more important in less mature language combinations… but managers find enormous linguistic barriers to dive into what’s happening.
That’s when the multilingual translations for Jira Service Desk come in. Automated translations can easily give managers the ability to understand the flow of communication in any language, as well as the topics customers request help with. And in this case, it really doesn’t matter too much that an AI translation doesn’t capture every subtlety of the customer request. QA Managers want to know the subject and the general reason why a ticket wasn’t solved.
Conclusions and key recommendations
These 5 scenarios have many aspects in common, starting from the simple fact that Jira Service Desk translations have allowed them to augment their capacity and serve their external or internal customers in situations that were previously painful. Business decisors need to consider very carefully what their support teams can do when their capacity is augmented, as it will give them the opportunity to optimize team structures with better coverage for minoritarian or more expensive languages.
Bridging communication across linguistic support groups makes domain-based ownership emerge as a model for both agents and managers: it’s not important what you can speak, but how much you know and can actually do in collaboration with others. In consequence, multilingual support teams will regularly increase their productivity and quickly diminish their backlog. In turn, smaller backlogs will impact overall support KPIs with shorter response times, more focus on harder issues, and a better ratio of open/resolved tickets.
Although each company has very specific needs, some recommendations can be applied to most uses cases. Customers should not be exposed to raw automated translations, which can be used as accelerators for internal communication across support levels and languages. Using a hybrid model where the final output is edited and proofread by a native speaker is a strong candidate for a generally accepted good practice (or linguistic hygiene, if you prefer).
Finally, measuring language barriers and their effect on support case resolutions will help predict the impact of translating customer emails and Jira issues. IT or Help Center manager will be able to move the needles of First Contact Resolution and time to resolution: don’t underestimate the impact of better communication.
Was this helpful?
Thanks!
![Capi [resolution]](https://avatar-management--avatars.us-west-2.prod.public.atl-paas.net/5c58630670c88b3ac427b731/9af058b3-54a7-4ac1-8e1f-fc30de903e63/128){kind=avatar}
Capi [resolution]
About this author
Inbound Marketing | Thought Leadership
Resolution
Berlin, Germany
19 accepted answers
16 comments