NorwayTechnology and telecom2026 – ongoingVisma
M2 Tarkka expense claim validation
Visma reduced returned expense claims by 25% with AI-assisted validation
Problem
Finance teams manually checked expense claims against tax rules and company policies. Conventional rules could not adequately assess free-text expense descriptions.
Approach
Visma added M2 Tarkka to its travel and expense software. It combined deterministic checks with language models accessed through Amazon Bedrock. Customers configured checks as warnings or blocking errors.
Results
Takeaway
Visma combined contextual checks with established validation rules, reducing the claims sent back for correction.
Categories
- Size
- Large enterprise
- Industry
- Technology and telecom
- Business function
- Finance, risk and legal
- Innovation type
- Product and service innovation
- AI technology
- Generative AI
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
AWS reported that Visma added language-model checks to expense validation. Customers reduced returned claims by 25%. The page carries no date; the Internet Archive first saved it on August 25, 2026.
Full details are in the original sources.
