Skip to content
uraAdvisory

AI Solutions

AI that does the routine. People who sign off.

We build AI into finance work where it saves real time: reading documents, preparing books, drafting reports and answering questions. Every output is reviewed by a qualified person before it is used.

Discuss an AI projectSee the portal

Our rule

AI suggests. People decide.

  • Reviewed by an accountant
  • Never trained on your data
  • Logged from upload to export

What we build

Eight AI services for finance and operations

Each one starts from a real process in your company and ends with people still in control of the result.

  • Process automation

    Repetitive finance steps, such as matching payments, posting recurring entries and sending reminders, handled by rules and AI, with people approving the exceptions.

    ExampleBank payments matched to open invoices every morning.

  • Invoice processing

    Supplier invoices read, coded to a category and account, checked for VAT and queued for approval.

    ExamplePurchase invoices ready for the VAT book without manual entry.

  • OCR automation

    Text pulled from scans, photos and PDFs, including receipts photographed on a phone, in Albanian and English.

    ExampleFuel and travel receipts read from a phone photo.

  • AI copilots

    Assistants inside the tools your finance team already uses, drafting reconciliations, variance comments and replies for a person to check.

    ExampleA first draft of the commentary on monthly variances.

  • AI business assistants

    An assistant that answers questions about your own figures, filings and deadlines, shows its sources and hands over to a person when it is unsure.

    ExampleWhat VAT do we owe this month, and why?

  • Reporting automation

    Management reports, KPI dashboards and board packs built from the ledger on a schedule instead of by hand.

    ExampleA monthly pack ready the day the books close.

  • Workflow optimisation

    A review of how documents, approvals and payments move through your company, then a redesign that removes waiting and rework.

    ExampleApprovals moved from email chains to a single queue.

  • Document intelligence

    Contracts, statements and letters classified and searchable, with key dates, amounts and obligations pulled out.

    ExampleRenewal dates extracted into the contract register.

How a document moves

Six steps from upload to export, one of them always human

The same pipeline runs in the client portal for every invoice and receipt. AI does the reading. An accountant decides what goes into your books.

  1. Upload

    PDF, JPG, PNG, Excel or Word, through the client portal.

    System step

  2. Read

    Text and layout recognised, including photos and scans.

    AI step

  3. Extract

    Supplier, customer, number, date, VAT, total, currency and terms.

    AI step

  4. Categorise

    A category and suggested account, with the VAT checked against 18% and 8%.

    AI step

  5. Review

    An accountant approves, corrects or rejects every reading.

    Human step

  6. Export

    Approved data flows into the books and the ATK VAT purchase and sales books.

    System step

Demonstration

See a sample invoice being read

This is the reading step as it appears in the portal, with a sample invoice. The extracted fields, the category and the VAT check all wait for an accountant before they are used.

Example Supplies L.L.C.

Office supplies

Invoice

INV-2026-0418

Bill to
Sample Company L.L.C.
Date
28 Oct 2026
Due
27 Nov 2026
DescriptionQtyAmount
Printer toner cartridges10650.00
Copy paper A4, box of 5 reams20360.00
Desk organisers12240.00
Net
1,250.00
VAT 18%
225.00
Total
€1,475.00

Payment terms: 30 days

Extracted fields

Sent for review
Supplier
Example Supplies L.L.C.
Customer
Sample Company L.L.C.
Invoice number
INV-2026-0418
Date
28 Oct 2026
Net amount
€1,250.00
VAT
€225.00 · 18%
Total
€1,475.00
Currency
EUR
Payment terms
30 days

Category

Administrative expenses

Suggested account: 6200 Administrative expenses

VAT check

225.00 is 18% of 1,250.00, the standard rate.

Accountant reviewIn review

Waiting for review before it reaches your books.

Demonstration with a sample invoice

Use cases

Where AI helps, by finance function

  • Accounts payable

    • Invoice capture and coding
    • Duplicate invoice detection
    • Payment run preparation

  • Accounts receivable

    • Matching sales invoices to bank receipts
    • Reminder drafts for overdue invoices
    • Customer balance summaries

  • Payroll

    • Checks of wage tax and pension calculations
    • Payslip and declaration preparation
    • Leave and contract date tracking

  • Tax compliance

    • VAT books prepared from approved documents
    • Rate and threshold checks
    • Deadline tracking and reminders

  • Reporting and FP&A

    • Monthly management packs
    • Draft variance commentary
    • Budget against actual dashboards

  • Contracts and documents

    • Renewal and notice date extraction
    • Obligation tracking
    • Search across contracts and letters

Governance

Rules we apply to every AI system

AI in finance is only useful if you can trust it and explain it. These rules apply to our own portal and to everything we build for clients.

  • People review AI output

    Every reading, category and draft is checked by a qualified person before it reaches your books, a filing or a third party.

  • No training on client data

    Your documents are processed only to deliver your service. They are never used to train AI models.

  • GDPR and Kosovo data protection

    Processing follows the GDPR and Kosovo Law No. 06/L-082 on the Protection of Personal Data, with processors bound by contract.

  • Audit trail

    Uploads, readings, corrections and approvals are logged with who did what and when.

  • EU AI Act awareness

    Kosovo is not an EU member, but many clients sell into the EU. We follow the EU AI Act approach: document what each system does, keep people in control and say clearly when content is produced by AI.

Engagement

How an AI project runs

Projects start small, prove themselves on your own documents and grow only when the results hold up.

  1. Map

    We map the process as it runs today, measure the volumes and agree what success looks like.

  2. Prototype

    We build a working version on a sample of your real documents and show you the results, errors included.

  3. Pilot with review

    The system runs alongside your team, with every output reviewed, until accuracy meets the agreed standard.

  4. Roll out and measure

    We hand over, train your people, document the system and review its performance every quarter.

AI automation projects start from €2,500.

See pricing

Find out what AI could take off your team

Tell us which process takes the most time. We will tell you honestly whether AI can help, and what it would cost.

Book a consultationSee pricing