Guideless Raises €1M Pre-Seed to Turn Software Workflows Into AI Training Guides

Lithuanian AI startup Guideless has raised €1 million in pre-seed funding as it looks to rethink one of the less glamorous but persistent problems inside software-driven organizations: documenting how work actually gets done.

The round was led by Superhero Capital, with participation from FIRSTPICK VC and angel investors Thomas Plantenga, Group CEO of Vinted; Vytautas Atkočaitis, CEO of Vinted Go; and Vinted co-founder Mantas Mikuckas.

Founded by Evaldas Bieliūnas and Dovydas Remeika, Guideless is building an AI-powered platform that turns software workflows into narrated training guides and structured documentation. The company says it now has more than 3,000 users and paying customers across 15 markets, with its technology in use by teams at Vinted, Masan Consumer Holdings and the Nevada Hospital Association.

The new capital will go toward product development, hiring and expansion across the United States, United Kingdom and Europe.

Turning software activity into structured training

Traditional software training frequently starts with someone recording their screen, explaining a process, editing the footage, adding captions and then repeating much of the work when the underlying software changes.

Guideless takes a different approach. Rather than treating the workflow as a single video recording, its Chrome extension captures the actions a user performs inside a browser and turns the important actions into separate, editable steps. The resulting guide can then be refined without having to reproduce the entire workflow.

Once a workflow has been captured, AI generates written instructions and narration for the individual steps. Users can reorder or remove steps, change screenshots, edit the transcript, adjust zoom and cursor highlights, and regenerate voiceovers as the underlying process changes.

That distinction matters because conventional training videos tend to become obsolete as soon as a user interface or workflow changes. By separating a process into editable components, Guideless is attempting to make software documentation behave more like a living knowledge asset than a finished piece of media.

The platform currently supports AI-generated text and voice selection across 43 languages. Guides can be distributed through links or live embeds, or exported in formats including MP4, PDF and Markdown. Guideless also includes analytics for tracking views, completions and where viewers abandon a guide.

For larger organizations, the company has also added shared workspaces and SAML 2.0 single sign-on, reflecting its push beyond individual tutorial creation toward enterprise knowledge management.

Building on experience inside Vinted

Guideless emerged in part from Bieliūnas’ experience working at Vinted Go, where he spent years involved in building and scaling the logistics arm of European second-hand marketplace Vinted.

Bieliūnas said the recurring difficulty of training employees across countries helped expose how much institutional time could disappear into explaining software workflows and rebuilding documentation.

“Leading commercial teams and onboarding employees across multiple countries for over a decade, including at Vinted Go, I saw firsthand how much time and money gets wasted on inefficient workflow training,” he said.

That history is reflected in the investors backing the round. Three of Guideless’ angel investors come from Vinted, while the marketplace itself became an early user of the software.

Vinted Group CEO Thomas Plantenga said the company had previously spent weeks rebuilding guides following user interface changes, whereas Guideless allowed those updates to be handled with significantly less manual work.

The startup soft-launched its product in late 2025 before formally incorporating in Lithuania in June 2026.

Moving beyond AI-generated training videos

The more consequential part of Guideless’ strategy is its attempt to move beyond generating tutorials.

Companies increasingly want AI agents to perform tasks across customer relationship management systems, enterprise resource planning platforms, support tools and other internal applications. Giving an agent access to those systems, however, does not automatically give it an accurate understanding of how a particular organization expects work to be performed.

A company’s real operating procedures often live across employee experience, informal instructions, outdated documentation and knowledge held by a relatively small number of people.

Guideless is betting that capturing workflows directly as employees perform them can create a more structured representation of that operational knowledge.

Because each workflow is stored as a series of discrete steps rather than a fixed recording, those processes could eventually provide context not only for human training but also for AI systems trying to understand how tasks should be completed inside a particular organization.

This shifts the potential value of the technology from answering “How do I use this software?” toward a more complicated question: “How does this company use this software to perform this specific job?”

Why operational context could become more valuable in the agentic AI era

The rise of autonomous AI could make that distinction increasingly important.

Large language models already provide broad knowledge about widely used business applications, but enterprise workflows are rarely generic. Companies customize software, create internal approval processes, establish exceptions and combine multiple applications into procedures that may exist nowhere outside the organization itself.

For AI agents to become reliable participants in those processes, they will need access to current operational context rather than relying solely on generalized model knowledge.

Guideless’ approach potentially creates one method of generating that context continuously: observe a real workflow, separate it into structured steps and keep those steps editable as the process evolves.

There are still substantial challenges ahead. Capturing a workflow does not automatically establish that it is the correct workflow, and organizations will need mechanisms for approvals, versioning, permissions and determining which procedures AI systems should be allowed to follow. Maintaining accurate operational knowledge at enterprise scale is also a much broader challenge than creating training materials.

But the €1 million round gives Guideless an opportunity to test whether a product that begins with a relatively straightforward use case, making software documentation faster to create and maintain, can evolve into infrastructure for how both employees and AI agents learn the way an organization operates.

If that transition succeeds, the longer-term opportunity may have less to do with replacing screen-recording software and more to do with building a continuously updated memory of how work gets done.