30 %
fewer email inquiries
More customers could find the answer through self-service.
ANONYMISED CUSTOMER CASE · 2 MILLION ANNUAL INQUIRIES
From operational function to strategic key player
Better knowledge gave 30 percent fewer emails, four times faster onboarding and over 300 percent higher efficiency with AI.
The organisation did not start with AI. It started by making its knowledge better and easier for employees to use. Customers were then given access to the same quality-assured knowledge. Only when that foundation was in place was AI connected.
A concrete meeting about your knowledge, workflows and realistic potential – not a standard demo.
3 months → 3 weeks
onboarding
New employees became confident and operational faster.
Over 300 %
higher efficiency
The effect varied between departments and case types.
22 → 0 days
response time in a selected queue
The backlog was cleared during the AI programme.
The results were achieved in different phases of the organisation's development with knowledge, self-service and AI.
Customer service got a seat at the table
The customer director describes how the work has helped move customer service from a traditional operational position to a seat at the table when the organisation's further development with knowledge and AI is decided.
Customer service already held much of the knowledge that the organisation's AI would use. Once that knowledge was structured, governed and made usable, customer service was not merely a receiver of new AI solutions. Customer service became a decisive prerequisite for AI to create real value.
01
Knowledge for employees
With around 2 million customer inquiries a year, even small improvements become noticeable. The organisation did not lack information. The challenge was ensuring that employees could quickly find the right, up-to-date and approved knowledge and apply it confidently in the customer interaction.
Therefore the organisation consolidated and structured its knowledge in Responza. The goal was not just a new system, but a knowledge foundation that worked in the real working day. It required clear roles, professional ownership and a way of working where subject matter experts could maintain the content themselves without waiting for IT or external consultants.
Onboarding was reduced from around three months to three weeks.
02
Knowledge for customers
Once the knowledge foundation worked for employees, the same quality-assured knowledge was made available to customers through self-service.
More customers could now find a reliable and understandable answer without first contacting customer service. That gave 30 percent fewer email inquiries and up to 40 percent shorter handling time.
It was not about removing the human contact, but about using it correctly.
03
AI on top of knowledge that worked
When the organisation later wanted to explore the potential of generative AI, it did not have to start from scratch. AI could be built on top of the knowledge, the roles and the governance that already worked in daily operations.
The organisation tested two AI assistants in selected written queues. One helped employees summarise and understand inquiries and relevant knowledge. The other drafted reply proposals for customer cases.
AI was tested where the balance between effect, quality and risk made the most sense.
From 14 to 59 cases a day
The effect varied between departments and case types, but in the affected areas efficiency improved by more than 300 percent.
In one concrete programme, production rose from 14 to 59 processed cases a day. A project group equivalent to around three full-time resources meanwhile handled 1,315 cases in the delimited trial period. In a selected inquiry queue, response time fell from 22 days to 0.
The results were not only about producing faster. Employees could get from the customer's inquiry to a qualified answer faster and spend more time on the professional assessment.
14
Before
59
After
processed cases a day
Do you want to know where the first documentable effect lies for you?
AI made different employees better in different ways
An important lesson was that AI did not only create value for one particular type of employee.
The experienced employee
A strong first draft and more time for complex, professional assessments.
The newer employee
Greater confidence and a faster path to a usable answer.
The digitally strong employee
Help with structure, language and precision.
Employees did not experience AI as a threat, but as a capable written assistant that took on part of the heavy writing work and made it easier to get off to a good start.
AI could help find, gather and formulate. The employee still had to understand the customer's situation, make the professional assessment and ensure the answer matched what the customer had actually written and experienced.
Higher pace without compromising on quality
The evaluation pointed to more consistent answers, better linguistic quality and greater uniformity in the organisation's desired answer style.
When an AI draft was not good enough, the cause could often be found in the knowledge foundation. The knowledge unit received concrete input to clarify and improve the content, after which the AI answers also improved.
AI did not just become another user of the organisation's knowledge. It also became an engine for discovering where knowledge could be improved.
AI works when knowledge works
The organisation's results were not created by putting AI on top of unstructured information.
- 01First, knowledge was made usable for employees.
- 02Then customers were given access to the same quality-assured knowledge.
- 03Finally, AI was put on top of a foundation that already worked.
That gave fewer inquiries, shorter handling time, significantly faster onboarding, higher capacity and employees who experienced AI as support in their daily work.
The technology was important. But it was the strong knowledge foundation that made the results possible.
How big is your potential with knowledge and AI?
If you are considering AI in customer service, the first question is not necessarily which model or chatbot to choose.
The first question is whether your knowledge is good enough for employees, customers and AI alike to trust it.
At a clarification meeting, we look at your current knowledge, workflows and the areas where you can realistically create the first documentable effect.
No standard presentation. We start from your reality.