A published article, a founder's account and a completed customer order each tell us something different about business automation. Treating them as interchangeable makes AI claims difficult to assess. A Lithuanian entrepreneur's projects offer a useful way to separate what can be observed, what the owner reports and what would require further evidence before drawing a commercial conclusion.
When a business says that agents handle its work, the natural follow-up is: which work, and how would we know? An agent drafting an email, sending it and helping resolve the resulting enquiry are separate achievements. Each needs a different form of evidence. A convincing demonstration of the first cannot establish the last on its own.
Kristin Zapasnik, founder of Verslomatika and a specialist in AI integration and business process automation, reports using agents across supplier research, partner communication, calls, customer email, stock monitoring, replenishment orders, delivery coordination, accounting-related tasks, advertising, social media, programming and website maintenance. The breadth of his account raises a useful question: what would a reader look for to understand each part of that work?
First, establish what the founder is actually claiming
Kristianas says that he chooses a business within his Verslomatika account and speaks to its coordinator. Each business has its own prepared specialist agents, knowledge and tools. The human sets the plan and priorities; the coordinator distributes work; specialists execute within the scope of their integrations and access. That is a more specific description than a general assertion that a company runs itself.
The owner's role makes the division of work clearer: commercial direction comes from the human, while the coordinator allocates tasks. For a reader, the useful next detail is which decisions and actions belong to each activity. A supplier search, a draft reply and an authorised external action would have different boundaries. Naming them makes the extent of the automation easier to follow.

An observable output answers a narrower question
One concrete public example is the Lithuanian-language Christmas guide to hand-casting kits on Topdovanos. It compares kits for couples and families, includes videos, explains preparation and use, and answers questions. Topdovanos is a portal for gift ideas and related guides. Readers can inspect how the guide organises the choices and whether its explanations help them understand the activity.
Christmas gifts for couples are easier to compare when the guide explains what is included and how the shared activity works.
The documented content process includes AI-assisted research, writing, structure, metadata and links, followed by editorial review, scheduled publication on the website and technical search optimisation. A reader can consider each stage separately: what information went into a draft, how it was reviewed and which version reached the website. Those questions make it easier to follow the work from preparation to publication.
Operational evidence follows a complete piece of work
The physical product business is Memory Casting, which offers kits for couples and families to make hand sculptures together. The owner reports one employee in his gift business. People and partners remain involved in physical tasks such as packing and delivery. An account of the digital work should therefore explain how it connects to those real-world activities.
To assess a particular operating claim, the useful evidence would follow the relevant activity. A stock alert, for instance, would need a clear data source and status; a replenishment order would need evidence of the action and its actual confirmation. The next question is what happened after the alert: was the information checked, who made the decision and what confirmed the resulting action? Following those steps makes the operating claim understandable.
- Account: What does the owner say the agents do, and within which boundaries?
- Output: What finished item can a reader actually inspect?
- Operating trail: What records would show that the relevant action and follow-up occurred?
- Outcome: What evidence would connect that work to a defined business result?

Commercial outcomes need their own evidence
A business result requires a defined measure and a meaningful period of observation. Producing more content is an output measure. A qualified enquiry or completed sale is a different event. To examine an improvement, readers would need to know what changed, over which period and which other factors may have contributed. That connects the operating account to a specific commercial question rather than treating all activity as the same kind of progress.
The same discipline applies to expansion. Kristianas plans to expand into France, Germany, the UK, Latvia, Estonia, Spain and Poland. For a British reader following the proposed UK expansion, useful future milestones would include a confirmed local offer, an operating customer journey and actual enquiries. These would show how the plan develops into a functioning market presence.

A useful claim is one that can be examined in parts
Verslomatika's business process automation work is described as integrating agents into actual businesses, potentially including process analysis, preparation and calibration, integrations, testing and maintenance. For a prospective user, that description suggests practical questions: which process is included, what information does it need, what can the agent actually do, and how is the result checked? Those questions make a discussion of automation concrete.
All three featured projects have the same owner. That relationship helps readers place the portal's coverage of the gift business in context. The practical next step is to ask one precise question at a time: what was produced, what happened next and which outcome followed? Applying those questions to other accounts of business AI makes their achievements and remaining ambitions easier to understand.
Comments