A short, honest record of AI we've actually put into production, not demos, not slideware. We keep client names private, so these are anonymised, but the systems and the results are real. A few here today; more as we're cleared to share them.
Each of these went past the demo and into daily use, wired to real data, with a person still accountable for it.
Toxic chat was driving players off a gaming platform while regulators circled. Context-aware moderation across 30+ languages told banter from abuse, cutting toxicity 45%, lifting new-player retention 20%, and dropping false blocks 75%.
Read the case → Credit scoring · AgricultureApproving farm credit meant sending a surveyor out for weeks. Computer vision on satellite imagery scored crop health and collateral, taking the company from dozens of applications a week to 25,000 farms a day and cutting underwriting cost over 70%.
Read the case → Data anonymization · HealthcareA medtech firm's clinical data was locked away by privacy law. An AI de-identification engine masked the patient details but kept the clinical signal, freeing terabytes safely and cutting the development cycle about 60%, at full HIPAA compliance.
Read the case → AI strategy & roadmap · RetailA retailer was burning budget on scattered AI projects. We ranked every idea by value and effort, cut dozens to three, and delivered a 90-day roadmap, with a quick win that proved a return in about eight weeks.
Read the case → Trust & safety · E-commerceBots and paid actors flooded a marketplace with fake reviews and harmful content. An anomaly engine plus human review blocked 95%+ before shoppers saw it, cut moderation cost sharply, and rebuilt trust.
Read the case → Search relevance · TravelA booking platform's search buried the best matches and its AI invented facilities that were never there. Human-verified data, relevance tuning and fact-check agents lifted top-result relevance 35% and cut invented-info reports 80%.
Read the case → Reporting automation · AccountingA firm that assembled dozens of clients' monthly reports by hand (from email, WhatsApp and photographed receipts) now reviews AI-drafted reports instead. Accurate, source-linked, and signed off by a professional. Live across dozens of clients.
Read the case → Support agent · AviationA large US airline was losing customers to slow email support. An AI agent now replies in minutes, holds anything touching money for a human, and cuts time-to-full-resolution by around 70%.
Read the case → Data readiness · InsuranceA leading Israeli insurance group's AI kept hallucinating. We traced it to a ~30% mismatch in the data, rebuilt the pipeline, and put a hard 85% confidence floor under every answer.
Read the case → Workflow automation · PropertyA large Israeli property manager had 8 people re-keying data by hand. We automated it end to end around a 15-year-old ERP. Average handling time fell from 4 hours to 30 seconds.
Read the case → AI adoption · RetailA large Western-European retailer had rolled out AI and almost no one used it. Role-specific, hands-on enablement took daily adoption from 15% to 85% and cut admin time by around 40%.
Read the case → Demand forecasting · RetailA large US retailer was losing money to spoilage and empty shelves. Per-branch demand forecasting a buyer approves cut waste ~30% and out-of-stocks 60%, and dropped buyer admin from 4 hours a day to 15 minutes.
Read the case → Dynamic pricing · RetailAn Israeli retailer was running promotions that quietly lost money. Real-time margin costing with a hard 12% floor blocks any loss-making deal before it hits the till, lifting gross margin by around 4%.
Read the case → Ops assistant · RetailStore managers waited days on HQ for a spreadsheet. An AI assistant in WhatsApp answers data questions in seconds and flags issues proactively, cutting report requests to HQ by ~70%.
Read the case → Fraud detection · InsuranceA European insurer was bleeding money to claims fraud while honest customers waited days. A real-time risk engine cut fraud losses 60% and false alarms 80%, and dropped legitimate approvals from 5 days to 30 seconds.
Read the case → Churn prediction · MediaA large Israeli content provider was losing thousands of subscribers a month with no warning. Early churn prediction with automated, personalised retention cut monthly churn ~25% and lifted usage 40%.
Read the case → Archive search · MediaA large US content company burned thousands of hours tagging and searching its archive. An AI content pipeline plus free-text search cut retrieval from days to seconds and localisation costs ~60%.
Read the case → Yield optimization · AdvertisingA leading Israeli ad agency sold inventory off fixed rate cards, losing at both ends. A yield engine with privacy-safe audience matching lifted ad revenue ~18% and cut proposals from days to minutes.
Read the case → Student retention · Higher edA veteran Israeli university was losing hundreds of first-year students to silent dropout. Early-warning student profiles flagged risk weeks ahead and cut first-year dropout by 25%.
Read the case → Smart facilities · Higher edA large university burned energy on empty halls and lost research to lab failures. Smart scheduling and predictive maintenance cut energy and upkeep ~18%, with zero lab downtime.
Read the case → Fraud detection · Public sectorA government body was losing public money to fraud while citizens waited months. A real-time detection engine cut fraud losses ~50% and false alarms 80%, and dropped approvals from weeks to minutes.
Read the case → Process automation · Public sectorA European authority took 45 days to approve permits across four disconnected systems. A privacy-safe AI environment with document extraction cut turnaround to 48 hours and clerical work 70%.
Read the case → Bid pricing · ConstructionA large Israeli infrastructure contractor kept winning tenders that lost money on site. AI bid pricing against its real field costs, with a hard margin floor, added ~3.5% gross margin and cut pricing from 3 weeks to 48 hours.
Read the case → Schedule control · ConstructionA civil contractor ran 35 infrastructure sites with the office always a few days behind the field. AI that predicts bottlenecks weeks ahead cut schedule overruns 35% and largely ended delay penalties.
Read the case → Trade promotion · Consumer goodsA leading Israeli food manufacturer hit every volume target while operating profit eroded on gut-feel promotions. AI promo simulation with a margin floor lifted gross profit ~3% and cut promo planning from weeks to minutes.
Read the case → SKU rationalisation · Consumer goodsA consumer-goods maker was choking on 1,500+ SKUs while 20% drove the profit. AI cost-to-serve analysis cut the catalog 25% with no revenue lost and 35% less production-line downtime.
Read the case → KYC & underwriting · BankingA large bank lost business clients to a 3-week credit-and-onboarding process while false alerts buried compliance. AI document abstraction and a green lane cut approvals to 24 hours and false alerts 70%.
Read the case → Fleet routing · LogisticsA heavy-haul fleet was dispatched by phone and gut feel, losing money to queues and empty return trips. AI routing with smart backhaul matching cut queue time 30%, empty miles 20%, and ran 15% more loads a day.
Read the case → Agent assist · TelecomA large telecom's agents put customers on hold to dig through five systems, and CSAT was sliding. A grounded, on-screen AI assistant cut in-call search 42% and lifted CSAT from 56% to 77%.
Read the case → After-call work · FinanceA financial firm's agents burned a third of the day on after-call summaries and CRM entry. AI auto-summaries approved in a glance cut average handle time up to 49% and lifted capacity 35%.
Read the case → Live translation · E-commerceA global e-commerce firm couldn't staff every language, so foreign-language tickets waited days. Secure in-workspace translation let every agent serve 20+ languages, cutting costs double-digit and lifting international CSAT ~35%.
Read the case → Ticket triage · DeliveryA delivery platform's agents juggled four systems to resolve one late order. AI triage with a ready credit and a hard guardrail cut handle time 15%, lifted first-contact resolution 30%, and handled 1.5× the tickets.
Read the case → Agent training · FintechA fintech platform's tickets doubled in days, but training an agent took a month. An AI training simulator and a grounded copilot cut onboarding to 10 days and scaled the desk 30% with zero compliance breaches.
Read the case → Omnichannel · E-commerceA global e-commerce firm ran a separate team per channel and language, so a customer who moved from email to chat had to repeat everything. One 360 workspace cut service costs 25% and lifted retention 40%.
Read the case → Lead scoring · FintechA global fintech's B2B reps drowned in unsorted leads while enterprise deals waited days and went cold. AI enrichment and predictive scoring lifted conversion 47.6% and cut strategic-lead response to 5 minutes.
Read the case → Sales copilot · SaaSA software platform took 3 months to ramp a rep who still froze mid-call. A real-time sales copilot with battle cards and an AI trainer cut ramp to 3 weeks and lifted deals per rep 150%.
Read the case → Customer success · SaaSA SaaS provider was losing B2B accounts at renewal because success only called at the last minute. AI early-warning six months out lifted the renewal win rate to 68.5%, cut churn 30%, and grew expansion 18%.
Read the case → Workforce planning · TravelAn airline's contact centres were overstaffed at quiet hours and swamped at peaks. AI demand forecasting and dynamic scheduling cut labour cost 18% and peak call abandonment 40%.
Read the case → Speech analytics · TelecomA telecom analysed just 1% of its millions of monthly calls, blind to why subscribers churned. AI that reads 100% of interactions caught failures in hours, cut repeat contacts 25%, and lifted NPS/CSAT 30%.
Read the case → Agentic automation · InsuranceAn insurer's chatbots could only answer, so 80% of digital requests still hit a human. Agentic AI wired into the core systems (verify, update, resolve) lifted self-service from 15% to 65% and cut operating costs 30%.
Read the case →We don't publish client names, not because we can't, but because the firms we work with would rather their competitors didn't know exactly what they've automated. The engagements are real and the numbers are theirs; the anonymity is deliberate.
If you want to go deeper than a web page allows, that's a conversation. We're happy to talk through how any of this was built, and what it would take on your own systems, on a call.
Reef TRH