Tallink: From Manual Email Routing to Intelligent CRM Automation
Tallink is one of the leading passenger and cargo transportation providers in the Baltic Sea region, operating across shipping, cargo, hotels, restaurants, retail, and e-commerce. With B2B enquiries arriving from multiple markets, in several languages and through different channels, managing customer requests efficiently is essential.
In 2024, Tallink set out to build a new CRM environment on SAP Sales and Service Cloud V2 together with Pearl Finland. The agreement was signed in early autumn 2024, and the first users started working in the new solution in spring 2025.
What started as a CRM initiative for B2B sales has since evolved into something broader: a more intelligent way of managing customer interactions, distributing work, and understanding sales activity across the organisation.
Starting point: A growing volume of enquiries, too much manual work
Before the new solution, Tallink's sales agents handled customer enquiries across multiple systems. Emails came from Finland, Estonia and from international markets in several languages, with requests split between Outlook and JIRA. The process was fragmented and routing enquiries to the right person was largely manual.
Managers spent time reviewing emails and distributing cases, relying on spreadsheets and information about working hours, vacations and workloads.
"We didn't have a proper routing system. One of the main goals was to allocate the work better and more efficiently: move from manual ticket management to intelligent, automated routing while creating a clearer picture of customer demand, agent workload and sales opportunities," says Jane Kohtring, Senior Product Manager, Tallink Group.
AI turned unstructured emails into structured sales data
Together, Tallink and Pearl Finland built a solution on SAP Sales and Service Cloud V2, covering key CRM processes such as case and opportunity management and contact handling.
At the heart of the new process is SAP AI Core, which acts as the intelligence layer between incoming customer emails and Sales and Service Cloud. AI reads each free-text email and automatically identifies key booking and sales information, including intent, language, route, date, guest count and special requests.
"The main thing that changed is that [agents] have a lot of information automatically in the system about the customer and about the email itself. The human agent still reads the original email and confirms that the extracted information is correct. So, AI does not replace human decision making, but removes repetitive work around the decision," Kohtring says.
Intelligent routing based on the right criteria
Another challenge was determining who should handle each opportunity. Tallink and Pearl Finland developed a rule-based routing model that considers factors such as language skills, availability, agent working time and vacation, current workload and the number of passengers in an agent's existing opportunities.
This reflects the fact that opportunities vary in complexity: a large group booking can require significantly more work than a simple enquiry.
"We are basically using four or five different criteria. We know what kind of sales opportunity takes a certain amount of time, and we can look at why something took so much time and how we can make the deal quicker. This gives managers a much clearer view of what is happening across the sales organisation," Kohtring says.
From one hour of manual work to more time for customers
With AI interpreting incoming emails and the routing model assigning them, the manual work of distributing enquiries has largely disappeared. Previously, managers relied on Excel sheets to track working hours, vacations and workloads when deciding how to assign new requests.
"Managers used to have Excel sheets showing working hours, vacations and workloads. That manual work, one hour a day, is now cut off thanks to AI. That time can now be used for customer conversations, solving complex cases and identifying additional sales opportunities, which is especially important during our high season in the summer," Kohtring says.
Future: expansions and fine-tuning
The initial rollout of AI-powered SAP Sales and Service Cloud V2 began with Estonian sales agents, followed by Finnish agents, where a larger volume of enquiries is handled. It is being expanded gradually as users become familiar with the new way of working and the routing logic is refined. Future plans include expanding the solution to additional B2B agents, hotels and B2C agents, as well as introducing further AI-powered capabilities for reading emails and automatically creating tasks.
"For us, the transformation is not about replacing human expertise. It is about making better use of it," Kohtring says.
As the solution continues to evolve, the ambition is clear: less time spent managing the work and more time spent creating value for customers.
