Gjensidige is one of the customers that taught me most about customer communication systems.
For readers outside Norway: Gjensidige is one of the large Nordic insurance companies, listed on the Oslo Stock Exchange, with operations in Norway, Sweden, and Denmark. General insurance is its core business, and in Norway it also offers pension and savings products.
XPER recently wrote about our cooperation with Gjensidige. It made me look back at what the work has taught me.
I have worked with Gjensidige on and off for many years. The work has covered document production, integrations, batch processing, online generation, print, email, SMS, archives, and careful modernization of systems that most people never see.
Customer communication systems are invisible when they work. A customer receives the right letter, with the right data and language, through the right channel. Nobody says, “Great batch architecture today.”
But when it fails, everybody notices.
A wrong amount in a letter. A missing document. A PDF that a screen reader cannot use. An SMS that bounces. A template that behaves differently in Denmark and Norway. A production run that creates thousands of wrong documents before anyone stops it.
That is CCM work. It may look boring from the outside, but it sits close to customer trust.
The work behind the templates
Document systems are often reduced to template design. In a large organization, the template is only the visible part.
Underneath you have business rules, customer data, product logic, legal text, language variants, channel selection, archive requirements, delivery systems, monitoring, fallback handling, performance constraints, and operational routines.
At Gjensidige, I worked with many of those layers. The projects included moving and modernizing document production, infrastructure around HP/OpenText Exstream, high-volume batches, online generation, local print, email, SMS, and bounce handling.
The requirement was always the same. The business had to trust every communication produced by the system.
Long-term customers are different
Long-term customer work is where I learn the domain.
You stop seeing only the current ticket. You understand why the system looks the way it does, which ten-year-old decisions still matter, which integrations are fragile, and which manual routines are protecting production from chaos.
You cannot buy that knowledge quickly. In the best XPER relationships, we become part of the customer’s technical memory. We know the systems, the people, the history, and what can happen when somebody changes the wrong thing casually.
Gjensidige takes the document domain seriously. The XPER article mentions trust, openness, and specialist competence. It sounds like normal company language, but in this case I think it is accurate.
The customer has to trust that you understand both the technology and the business risk. You also have to say when something is fragile, unclear, or more complicated than it looks.
Accessibility is the next serious chapter
Gjensidige is taking accessibility seriously. I respect that.
The European Accessibility Act can easily become another compliance deadline. Customer communication has a more basic test.
If a customer cannot read or navigate an important document, the communication failed. The batch status, visual PDF, and archive receipt do not change that.
The message did not reach the person properly.
This is especially serious in insurance, banking, healthcare, pensions, and public services. These documents explain rights, obligations, money, risk, decisions, and next steps.
Accessibility is part of the communication system doing its job.
AI around controlled communication
AI is changing the work around these systems. It can analyze templates, explain old logic, classify documents, simplify language, detect anomalies, generate tests, and support migrations. I use it for several of those jobs already.
It does not remove the need for controlled communication systems. It makes the boring foundations more important.
The output has to be correct. The audit trail matters. The business owns the message, and legal text cannot be improvised between runs.
I think AI is most useful around CCM, where it helps teams understand, test, migrate, and operate controlled systems.
That is why the Gjensidige work has stayed interesting. It combines old enterprise reality with accessibility, modernization, automation, and AI-assisted quality work.
The system may be invisible, but its output is part of the relationship between Gjensidige and its customers. A wrong document damages that relationship.