Your software keeps changing.
Your documents do not.

Khangul.ai checks whether your actual software still matches its design documents and manuals — and shows what has changed, what is different, and what is missing.

See how it works Talk to us about local validation

Design, system and manual drifting apart over time

DESIGN SYSTEM MANUAL

The problem

Do your software systems still match the original documents?

Rules change → Systems change → Documents fall behind → Risk grows

People are trained on the wrong information

New staff learn the system from manuals that no longer describe how it actually works.

Handover becomes difficult

A new team or a new vendor takes over a system that nobody can fully explain.

Reviews lack clear evidence

When an audit or a review asks why the system behaves this way, there is no clear evidence to show.

Policies change. Work changes. Software changes.
The documents do not change as fast.

Why now

AI makes software development faster. Verification is becoming the next bottleneck.

AI can help teams build and change software faster. But testing, verification and documentation still require significant manual work.

As software changes faster, it becomes harder to know what was changed, what was actually tested, and whether the documents still describe the system correctly.

Software can now change faster than organizations can verify and document it.

What we do

Four steps, from documents to evidence

Read the documents

We take in the design documents and manuals to understand what the system was originally meant to do.

Check the running software

We automatically walk through the actual screens and workflows that exist in the system today, and record what happens.

Find the gaps

We identify what is missing, what is different, and what is no longer consistent.

Produce the records

Because every step is recorded as it runs, the result is not only a list of problems — it is a set of documents your team can review and share.

What Khangul.ai produces

Findings report What does not match between the documents and the system, item by item.

Video recordings Each check is recorded as it runs, so anyone can see what actually happened.

Test results Which parts of the system were checked, and what passed or failed.

Draft operation manual A manual written from how the system actually works today, ready for your team to review.

For technical teams: Khangul.ai builds state-transition models of both the designed and the actual system behavior, and compares them.

Khangul.ai — product demo

Why it matters

Less manual work. Better control. More confidence.

For software & QA teams

Less manual QA work Reduce repetitive screen-by-screen checking and test execution.

Faster change verification When software changes, identify what changed and what needs to be checked without starting the review from scratch.

For PM, operations & delivery teams

Easier handover New staff or a new vendor can understand the current system with less dependence on the people who operated it before.

Documentation closer to reality Keep operational information aligned with the software people actually use.

For system owners & reviewers

Evidence, not just pass or fail See what was checked, what happened, and why an issue was identified.

More confidence in delivered software Check whether the system that was delivered still matches what was designed and agreed.

Deployment

It runs in your environment — not only in the cloud

Runs on a normal PCNo expensive GPU infrastructure is required to get started.

Your data stays insideDocuments and screens are not sent to an outside service.

Works offline and on-premiseDesigned for closed and restricted networks from the start.

No system replacementIt works on top of the systems you already have.

Public services

Why this matters for public services

Better governed software can support more reliable and sustainable digital services.

More reliable digital services Better verification helps reduce the risk that undocumented changes and outdated information affect service delivery.

Better continuity When staff or vendors change, critical knowledge about the system does not have to leave with them.

Greater accountability Clear findings and evidence help organizations understand what was delivered, what changed, and what needs to be corrected.

Stronger local IT capability Local software teams can adopt more structured verification and quality practices without depending on expensive cloud infrastructure.

Better governed software helps public organizations provide more reliable services to the people who depend on them.

Current status

Built in Korea. Ready for local validation.

Our core product is built and is currently being validated with software teams in Korea.

For international markets, we begin with one real software project to understand the local environment — including documents, workflows, language and deployment requirements.

This allows us to configure and localize the product for each environment together with local partners.

There is no need to replace existing systems. We start with a focused local validation, and if the problem and fit are confirmed, we define a small pilot together and measure the results.

How local validation starts

STEP 1Understand the local environmentLearn how the organization develops, operates and maintains its software.

STEP 2Review a real exampleReview selected design documents, manuals and actual software workflows from a real project.

STEP 3Define what needs to be localizedIdentify language, document, workflow and deployment requirements.

STEP 4Plan a local pilot togetherDefine the scope, expected benefits and how the results would be measured.

Why Khangul.ai

Built by experienced practitioners

Khangul.ai was founded in Korea in 2026 by two professionals with decades of experience in software engineering, public-sector IT and global technology projects.

Founder & CTO

Jin-Chul Park

Professional Engineer in Computer System Applications, a Korean national-level professional engineering qualification. Certified senior information systems auditor. More than 12 years working on public-sector and defense systems, and around 30 years in software engineering.

Co-founder & COO

Sung-Kwen Lee

Around 25 years of technology leadership at Samsung C&T, Cisco and Afiniti, delivering large-scale AI and digital transformation programs for major enterprises.

6 patent filings in Korea, covering the core verification pipeline and related technologies.
Core filing 10-2026-0032578 — state-based exploration and comparison of designed and actual software behavior.

Your software keeps changing.Your documents do not.