All posts

Orbio World: 3,000 invoices in two hours

Odeta Isevičiūtė5 min read
Orbio World: entering 3,000 invoices went from a week to two hours.

Two or three people, about a week of work, 3,000 invoices. After the ChangeAI program, Orbio World's finance team does the same job in roughly two hours with Claude Cowork.

Debt reconciliation now takes about 80% less time, and more than half the team has become AI enthusiasts. That number is still growing.

ProgramParticipantsPeriod
AI transformation program20 accounting and finance specialistsFebruary–March 2026

The problem: a system that can't do it all

Orbio World is a fast-growing international e-commerce company with more than 10 million customers in 190 countries. Its finance and accounting department serves several legal entities, works under different financial regulators and handles a huge flow of documents.

Because of its scale and structure, the company can't find a single finance system that covers every need. Even the XML payment files for the bank had to be prepared by hand, because the system doesn't generate them.

So the team was doing two jobs at once. It ran daily finance operations across several companies, marketplaces and regulators, and it patched the system's gaps by hand. There were never enough people, and the workload kept growing.

Why the finance team learned to do it themselves

Mantas Baliūtis, Head of AI at Orbio World, made a strategic call: the internal AI team works on solutions that create direct business value, and an external partner builds the rest of the company's AI skills. The goal was simple. Every team member becomes an active AI user. Orbio World chose ChangeAI for the job.

We ran the program for 20 specialists across three teams, structured around ChangeAI's AI iceberg:

  • Session 1. Introduction to AI, safe use, the core tools and their features, prompts and assistants. Theory alternated with demos and practice, so participants saw for themselves how the results change.
  • Homework between sessions. Every participant built their own AI assistants and kept notes on where time was being lost in their daily work. Applying it started straight away, in their own jobs.
  • Session 2. Hands-on work on the real Orbio World processes participants had captured between sessions. Claude Cowork became the main tool: the team used it to solve daily problems and automate repetitive, time-consuming work.

The program also pushed the teams to share knowledge and build their own learning habits, so the impact keeps growing after the training ends.

Results: a few examples out of many

The team finds new uses every week, so this table shows only part of what has already changed.

ProcessBefore the programAfter the program
Debt reconciliation (PDF preparation and attachments)2 people, 2 days a month1 person, a few hours of checking
XML payment files for the bankManual preparation, errors, delaysAutomated with Claude Cowork
Entering 3,000 invoices2–3 people, about a weekAbout 2 hours with Claude Cowork
Invoice data from emailsHanded to IT, waited 3+ monthsFinance team does it themselves
Analysis and data processing2–3 days per taskA few hours

The team built some of these during the training. It found the rest on its own, after the program.

"New ones turn up every week. Work nobody planned for, so more and more often I ask: have you tried solving this with Cowork yet?" Vestina, Head of Accounting, Orbio World

The AI team went back to revenue

One of the biggest changes: the finance team now handles work it used to hand over to the AI team.

Saving, renaming and archiving invoice data had been logged as a request to the AI team three months earlier. It sat in the queue because the AI team was building revenue-generating solutions. Now the finance team does it themselves, with Claude Cowork.

The gain is twofold. The finance team stops waiting and gets the result immediately. The AI team spends its time on systems that drive the company's revenue.

"My team has to build solutions that grow revenue directly. But every hour the finance team spends on manual work is also the company's money. ChangeAI let us solve both problems at once: they enabled the finance team, and we kept our focus where it's needed most." Mantas Baliūtis, Head of AI, Orbio World

What changed in the team's day-to-day

The program built skills, and it also changed how the team works every day:

  • part of the team uses AI in its daily work;
  • the accounting department set up an internal AI knowledge-sharing channel;
  • at regular stand-ups, colleagues share their wins;
  • other departments have started asking for the know-how;
  • a few colleagues use AI so heavily they have hit the daily usage limit.

A change like this needs internal AI ambassadors. At Orbio World, that was Vestina: she gets colleagues excited, helps them past the first hurdles and keeps looking for new uses. In every ChangeAI program we look for people like her and back them, because they are the ones who carry the change on after the program ends.

"The whole team uses it and really sees results. Colleagues tell each other what they managed to do with it, we share ideas and skills, we look for new tasks we could speed up, and because of that we find new ways to use it every day." Vestina, Head of Accounting, Orbio World

The key insight

"Only when you know 100% what result you need can you really master AI for handling those processes." Vestina, Head of Accounting, Orbio World

AI creates the most value when the process experts run it themselves: the people who work with the data every day and know best what result is needed.

The tasks solved during the program were the start. The lasting impact comes when the team has enough practical knowledge to apply AI to its daily processes on its own.

"The investment in the team's AI skills paid off instantly. Our finance processes got faster, and we freed up our AI team's resources to build profit-driving solutions. It's a double strategic win that makes our company more efficient every day." Martynas Nenėnas, CFO, Orbio World

Quotes translated from Lithuanian.

If your team handles a similar flow of documents, you'll find more results on our customers page, and more about the program for teams on the AI Sprint page.

Message us