Data readiness diagnostic
๐Ÿ”’ 100% local ๐Ÿ‘ Read-only ๐Ÿšซ Zero uploads ๐Ÿงฎ Deterministic

Is your data actually
ready for AI?

A local diagnostic tool that audits a company's files, data and folder structure โ€” typically around older ERP/CRM systems โ€” and answers one question: can this business migrate to a modern, AI-automated system?

"Before a company invests in AI or automation, we check whether its data can actually support it."
The diagnosis

Nine families of data problems โ€” counted, not guessed

It reads every Excel and CSV file, profiles it, and surfaces exactly what's wrong. Deep validation is tuned for Greek business data.

๐Ÿ“Š

Profiling

Rows, columns, empty cells and per-column missing % across every data file.

๐Ÿ‘ฅ

Exact & fuzzy duplicates

Identical rows, duplicate emails/phones/AFMs, plus similar-name clusters and the same contact under different names.

โœ‰๏ธ

Email validation

Format check on every email value in the dataset.

๐Ÿ“ž

Greek phones

Mobile (69โ€ฆ) and landline (2โ€ฆ) validation with +30 handling via the phonenumbers library.

๐Ÿงพ

Greek AFM / VAT

9-digit format plus the official checksum validation โ€” catches numbers that only look valid.

๐Ÿ”ค

Greek / Latin names

Mixed-script detection and Greeklish-aware matching: ฮ“ฮนฯŽฯฮณฮฟฯ‚ = Giorgos = Giwrgos.

๐Ÿท๏ธ

Category chaos

The same category written many ways โ€” case, spaces, accents, spelling โ€” collapsed and reported.

๐Ÿ’ฏ

Explainable score

A 0โ€“100 readiness score where every deducted point maps to a specific counted problem. No black box.

๐Ÿ› ๏ธ

Plain-language Fix Plan

Step-by-step remediation for non-technical readers โ€” who does what, which free tools, and time estimates that scale with the counts.

Built for confidential data

Nothing leaves the machine. Ever.

This is a diagnostic instrument, not an AI "magic cleaner" and not a SaaS. It never modifies, merges or rewrites your data โ€” and it never sends it anywhere.

๐Ÿ”’

100% local, localhost-only

The web UI binds to 127.0.0.1 โ€” unreachable from the network. No cloud APIs, no OpenAI / Claude / Gemini, no internet processing.

๐Ÿ‘

Read-only by design

It never modifies, overwrites or deletes input files. It reads, analyzes, and writes a report to a local output folder โ€” nothing else.

๐Ÿงฎ

Deterministic, not AI-guessed

All analysis is deterministic local Python. Same files in, same report out โ€” auditable and repeatable.

๐Ÿšซ

No telemetry, no tracking

No uploads, no analytics, no phone-home. Every input file is treated as confidential.

โš–๏ธ

GDPR authorization gate

The UI locks the scan button until both parties confirm the legal basis โ€” written authorization, NDA, Article 28 DPA, scope and local-only handling.

๐Ÿ‡ฌ๐Ÿ‡ท

Greek-law aware

GDPR data-protection findings tied to results, referencing GDPR (EU 2016/679) and Greek Law 4624/2019. A DPO checklist โ€” not legal advice.

Three steps, no upload

Point it at a folder. Read the verdict.

Runs as a Mac app or a command line. Double-click, a local server starts, your browser opens โ€” pick folders visually or type a path.

Pick the folder

Navigate to Home, Desktop, Documents, an external drive or a mounted Windows/NAS share (SMB). Choose a full scan or data-files-only.

Deterministic analysis

It audits every Excel/CSV, inventories all file types, hashes duplicates, detects the source ERP/CRM and scores migration readiness โ€” locally.

Readiness report + fix plan

A single HTML report opens in the browser: score, every problem found, GDPR findings, and a costed, step-by-step plan. Print to PDF for the client.

Beyond the score

From "what's wrong" to "how to fix it"

Full scans go past data quality into the migration itself โ€” which system produced the files, what a clean structure looks like, and a human-approved way to build it.

๐Ÿš€

Migration readiness (0โ€“100)

Can these files move to a modern system, and what blocks it? Factors: structured-data quality, legacy formats, duplicate files, version chaos, folder structure, stale content, permissions. Hard blockers listed separately.

๐Ÿ”Ž

Source-system detection

Deterministic signature matching recognizes likely origins โ€” SoftOne, Epsilon Net (Pylon/Extra), SingularLogic Galaxy/Control, Eurofasma, Entersoft, Atlantis, ฮšฮตฯ†ฮฌฮปฮฑฮนฮฟ, PRISMA Win, Clarion, Access apps, HubSpot/Salesforce exports โ€” with confidence, evidence and a vendor-specific migration note.

๐Ÿ—๏ธ

Transition Studio (human-approved)

Proposes a clean, numbered Greek-SME folder structure and routes files by precedence. A human approves every action group. Safe copy mode by default (originals untouched, checksum-verified); in-place replacement is gated behind a typed confirmation.

๐Ÿ—„๏ธ

Legacy formats & full inventory

Flags .dbf, .mdb/.accdb, .xls/.doc, .tps, .btr, .bak, .sql โ€” each with a conversion action. Inventories every file type, finds content-hash duplicate files (wasted disk), naming chaos ("final", "copy", "_v3") and stale files.

Simple and explainable

One score. Four grades. Zero mystery.

Start at 100. Subtract capped points per problem family โ€” missing data, duplicates, invalid identifiers, name conflicts, inconsistent categories, unclear columns. The report shows every deduction with its count.

85+
Ready
Good to build on
65โ€“84
Needs Work
Fixable gaps
40โ€“64
At Risk
Serious cleanup first
<40
Not Ready
Don't build yet
โš ๏ธ
Diagnostic only. The report is a technical data-quality assessment, not a compliance audit โ€” it does not replace legal or GDPR advice.
๐Ÿ”
Authorized use only. Real client data must only be processed after written authorization, a signed NDA and an appropriate GDPR data-processing agreement. Until then, use the included fake sample data. Treat every input file as confidential.
AIA Systems

Know before you build.

Spend money cleaning bad data, or spend it building on top of it. This tool tells you which โ€” before the transition, on the client's own machine.