AI Scanning FAQ

The problem: Typing bet details manually takes 2-5 minutes per bet and introduces transcription errors in stakes, odds, and returns — fields where even small mistakes distort your analytics.

The technology: Computer vision models like GPT-4 Vision are trained on millions of document images and can recognize text, numbers, logos, and layout patterns across thousands of different bookmaker formats — far more reliably than manual copying.

How Zort works: When you photograph a bet slip, it's securely sent to our multi-provider AI pipeline (GPT-4 Vision primary, Groq and Alibaba Cloud Vision as automatic fallbacks). The AI identifies stake amount, odds in any format, bookmaker, sport type, and potential returns. The entire process takes 2-3 seconds and achieves 98.5% accuracy.

The concern: Bettors understandably worry that automated scanning might misread critical values — a wrong odds digit or misplaced decimal point could corrupt their entire performance record.

The evidence: Based on analysis of over 50,000 scanned slips, Zort achieves 98.5% accuracy — meaning 98-99 out of every 100 bets have all fields correctly extracted on the first attempt. By comparison, manual data entry typically produces errors in 3-5% of entries due to typos and misreading.

The safeguard: Zort's multi-provider fallback system means if one AI struggles with a particular format, another handles it. Plus, every scan displays a confirmation screen where you can tap to review and edit any field before saving — so nothing is recorded without your approval.

The challenge: Bettors use dozens of different bookmakers — some in-shop with paper slips, others via apps with digital screenshots — each with unique layouts, fonts, colours, and formats.

Why AI solves this: Unlike rigid template-based scanners that break when formats change, AI vision models learn generalizable patterns — text regions, number formats, layout structures. This means they adapt to bookmakers they've never seen before.

What Zort supports: Physical paper slips (including crumpled, folded, or slightly damaged ones) from betting shops like William Hill, Paddy Power, and Coral. Digital screenshots from apps like Bet365, Betfair, DraftKings, and any other bookmaker worldwide. No bookmaker-specific setup required.

The problem: A complete bet record requires 8-10 data points (stake, odds, returns, bookmaker, sport, bet type, date, currency). Manually entering all of these for every bet is tedious, and missing even one field compromises your analytics.

Why automation matters: AI vision models extract all structured fields from a single image in one pass — no multi-step form filling, no switching between the bookmaker app and your tracker.

What Zort captures: Stake amount, odds (auto-converts between decimal, fractional, and American), potential returns/winnings, bookmaker name, sport and event details, bet type (single, accumulator, each-way), date and time of placement, and currency. Every field is editable in the confirmation screen before saving.

The problem: Manual bet entry takes 2-5 minutes per bet — opening the tracker, navigating to the right form, typing each value. Over 200 bets per year, that's 7-17 hours of tedious data entry. Most bettors give up and stop tracking consistently.

The speed comparison: Cloud-based AI processes an image, extracts structured data, and returns results in a single network round-trip — approximately 60x faster than manual entry.

Zort's performance: The complete process takes 2-3 seconds on average (image upload + AI processing + displaying results). Batch scanning 10 slips takes about 20-30 seconds with parallel processing. Over a year of betting, this saves 10+ hours compared to typing every bet manually.

The situation: Bettors aren't always online — at stadiums with poor signal, in betting shops, or travelling. But waiting to record bets means forgetting details or losing slips entirely.

The trade-off: Cloud-based AI delivers the highest accuracy (access to latest models), but offline recording is essential for capturing bets anywhere, anytime.

How Zort handles it: All your betting data is cached locally, so your full history and analytics work offline. If you're without connectivity, photograph slips and queue them — they'll process automatically when your connection returns. You never lose a bet slip, even without signal.

The concern: Bet slips contain personal financial information — stake amounts, bookmaker accounts, and betting patterns. Bettors rightfully want to know what happens to their images after scanning.

Our approach: Privacy by design means minimizing data retention. Since the AI only needs the image temporarily to extract text data, there's no reason to store the original photograph.

What Zort does: Images are processed by the AI to extract betting data, then immediately deleted from our systems. Only the extracted text data (stake, odds, bookmaker, etc.) is stored in your encrypted Zort account. Your bet slip images are never retained on our servers.

When errors happen: While 98.5% accuracy means errors are rare, they can occur with poor image quality (blurry photos, low lighting), unusual fonts, or very complex multi-leg accumulators with many selections.

Why it's not a problem: Zort is designed with a human-in-the-loop workflow. No data is saved without your explicit review — every scan shows a confirmation screen displaying all extracted values.

How to correct: Simply tap any field to edit it. Changes take effect instantly. Over time, the multi-provider AI system continuously improves, so formats that caused errors previously are handled better in future scans.

The challenge: There are hundreds of bookmakers worldwide, each with distinct slip designs, logos, layouts, and terminology. A scanning system must work across all of them without manual configuration.

How AI handles this: Rather than using fixed templates per bookmaker, Zort's AI learns generalizable patterns for how betting information is presented — so it adapts to formats it's never explicitly trained on.

Confirmed bookmakers: Bet365, William Hill, Paddy Power, Betfair, Ladbrokes, Coral, Sky Bet, 888sport, Unibet, BetVictor, Betway, DraftKings, FanDuel, BetMGM, PointsBet, Caesars, BetRivers, and hundreds more. If you use a bookmaker we haven't seen before, the AI still extracts the key information.

The situation: Bettors often accumulate multiple slips throughout a day at the bookmaker, or have a week's worth of screenshots to catch up on. Scanning one at a time would still be slow.

How parallel processing helps: Cloud-based AI can process multiple images simultaneously rather than sequentially, dramatically reducing total wait time for batch scans.

Zort's batch scanning: Photograph multiple bet slips in quick succession. Each is processed in parallel, so scanning 10 slips takes only about 20-30 seconds total — versus 20-50 minutes of manual entry for the same batch. First-image results appear while remaining slips continue processing in the background.