Ortem Technologies

    Innovation Lab

    The Ortem Innovation Lab

    Research, Prototype, Ship — AI Products Built to Solve Real Problems

    Where we research, prototype, and stress-test emerging technologies — then ship them as real products. Fleet Sync Pro, Azimuth, and Residenta all started here. Every client recommendation we make is backed by hands-on Lab experience.

    AI & ML Services

    Why an Innovation Lab?

    Clients come to us because we understand the difference between what is technically possible today and what is production-ready. The Innovation Lab is where that judgment is developed — through hands-on prototyping, benchmarking, and shipping real products. When we recommend a technology on a client project, it is because we have already built with it, broken it, and fixed it. Fleet Sync Pro, Azimuth, and Residenta are not prototypes — they are live products that validate our engineering approach at scale.

    Products Born in the Lab

    Five live SaaS products — each solving a real market problem with AI-first architecture, built and operated by the Ortem engineering team.

    Fleet Sync Pro logo
    Logistics & Fleet ManagementLive

    Fleet Sync Pro

    Modern Fleet Operations SaaS

    The Problem

    Fleet managers across logistics, transport, and field services were running operations on spreadsheets and disconnected tools — manually tracking oil changes, fuel fill-ups, parts inventory, and driver assignments across 100+ vehicles. Missed maintenance reminders caused costly breakdowns. Fuel spend was unmonitored, allowing 10–20% budget leakage. Driver-reported issues sat in email inboxes instead of triggering structured work orders. With no role-based access, managers, technicians, and drivers operated from the same admin views, creating confusion and compliance risk.

    What We Built

    Fleet Sync Pro is a multi-tenant SaaS platform that centralises every aspect of fleet operations — vehicle registry, work order management, maintenance reminders, fuel tracking, digital inspections, parts inventory, and role-based dashboards — into a single beautifully designed web application accessible from desktop, tablet, and mobile.

    8–15% fuel cost reduction40%+ downtime reduction3x faster work order throughput
    View Case Study How We Built It

    What We Solved

    • Replaced spreadsheet-based maintenance tracking with interval and mileage-based automatic service reminders
    • Built a work order lifecycle system (Open → In Progress → Pending → Closed) with auto-numbering and parts tracking
    • Created a real-time fuel tracking module with MPG calculation and 30-day cost trend analytics
    • Implemented 5-role RBAC (Admin, Fleet Manager, Maintenance Manager, Technician, Driver) at PostgreSQL RLS level
    • Delivered driver-facing issue reporting with photo attachments converting directly to work orders
    • Built multi-tenant architecture with per-organisation data isolation — zero leakage between fleet operators

    How AI Transforms This Product

    • Predictive maintenance AI analyses vehicle service history to forecast upcoming failures before they occur, reducing emergency breakdown costs by 30–50%
    • Anomaly detection on fuel fill-up patterns flags potential fuel theft or inefficient driving behaviour automatically
    • AI-driven route optimisation (v2 roadmap) will reduce fuel consumption by calculating optimal daily routes per driver
    • Natural language work order creation — technicians describe a job in plain English and AI creates a structured work order with suggested parts and labour estimates

    Why This Is the Best Solution

    Fleet Sync Pro is the only fleet management SaaS built with per-organisation Row-Level Security from day one — meaning your fleet data is isolated at the database layer, not just the application layer. Combined with a real-time operations dashboard, automated service reminders, and a five-role access system that mirrors how real fleet teams are structured, it replaces three separate tools (maintenance tracker, fuel log, work order system) with one coherent platform.

    Azimuth logo
    SaaS & Agency ToolsLive

    Azimuth

    AI-Powered Project Estimation & Proposal SaaS

    The Problem

    Software agencies and freelance development teams were spending 4–8 hours per proposal manually estimating project scope, writing requirements documents, building cost spreadsheets, and assembling client-facing proposals — generating no revenue and repeating from scratch for every new enquiry. Estimation accuracy was inconsistent, CRM data lived in separate tools requiring re-entry, and proposals looked unprofessional because there was no time for design. There was no learning mechanism, meaning the same estimation mistakes repeated indefinitely.

    What We Built

    Azimuth is an all-in-one AI estimation and proposal SaaS that takes raw client requirements and outputs a structured scope, cost breakdown, PRD, TRD, and client proposal — all in minutes. With a built-in CRM, branded email delivery, shareable document links, and a self-improving AI estimation engine, it replaces the manual proposal process that costs agencies 4–8 hours per deal.

    4–8 hrs saved per proposalPRD + TRD + Proposal in one click7-stage CRM pipeline built in
    View Case Study How We Built It

    What We Solved

    • Built an AI estimation engine that parses raw requirements into module-feature-task hierarchies with per-role hour estimates in under 5 minutes
    • Created one-click document generation for PRD, TRD, and Client Proposal via Anthropic Claude with section-level editing
    • Implemented version history on all documents — every save creates a snapshot with full rollback capability
    • Integrated a full Kanban CRM pipeline (7 stages) directly linked to projects, eliminating data re-entry between tools
    • Delivered branded email proposal delivery via Resend with open tracking — agencies know when clients read proposals
    • Built an AI-vs-actuals feedback loop that captures estimated vs. actual hours per task to improve future estimates

    How AI Transforms This Product

    • Anthropic Claude parses raw requirements into structured module-feature-task breakdowns with per-role hour estimates — replacing hours of manual analysis
    • Section-level AI regeneration lets users refine individual document sections without regenerating the entire document
    • Training feedback loop ingests actual vs. estimated hours per task after project completion, narrowing estimation error over time
    • AI financial analysis provides pipeline health insights — identifying which deal stages have the lowest conversion and why

    Why This Is the Best Solution

    Azimuth is the only project estimation tool that combines AI scope decomposition, three-document generation (PRD + TRD + Proposal), a built-in CRM pipeline, and branded email delivery in a single platform. Competitors handle estimation or proposals or CRM — Azimuth handles all three simultaneously. The self-improving AI feedback loop means the platform gets more accurate with every project, creating a compounding advantage for agencies that use it consistently.

    Residenta logo
    Real Estate & PropTechLive

    Residenta

    AI-Powered Property Management SaaS

    The Problem

    Property managers and landlords were operating with a fragmented toolset — tenant communication scattered across emails and calls, maintenance requests tracked in spreadsheets, rent collection manual and without visibility, and zero predictive insight into upcoming issues. As portfolios grew from 10 to 100+ units, manual processes collapsed: maintenance was missed, lease renewals expired without action, rent reminders required individual follow-up per tenant, and residents had no way to self-serve without calling the manager.

    What We Built

    Residenta is an AI-powered property management SaaS that centralises every aspect of property operations — properties, units, leases, maintenance, finances, notifications, and a resident portal — into one platform. AI handles ticket classification, maintenance prediction, lease assistance, and financial insights. Automation handles rent reminders, overdue invoices, and lease expiry alerts. Residents self-serve. Managers focus on decisions, not admin.

    60% faster maintenance resolution94%+ on-time rent collectionNear-zero lease renewal misses
    View Case Study How We Built It

    What We Solved

    • Built full lease lifecycle management including single and bulk lease renewal — process 50 expiring leases in one action
    • Created AI maintenance ticket classification that auto-assigns category and priority to every incoming request without manager triage
    • Implemented a resident self-service portal reducing inbound manager queries by 45%
    • Automated three daily cron jobs: overdue invoice marking, 30-day lease expiry alerts, and 3-day rent due reminders
    • Delivered real-time in-app notifications via Supabase Realtime — residents receive alerts the moment a ticket status changes
    • Built a context-aware AI Lease Assistant chatbot that knows tenant name, rent, deposit, and dates per lease

    How AI Transforms This Product

    • AI maintenance ticket classification auto-assigns category (plumbing, electrical, HVAC, structural) and priority to every incoming request in seconds — no manager triage needed
    • Predictive maintenance AI analyses ticket history per unit and forecasts upcoming issues with confidence scores and predicted dates, enabling scheduled repairs before failures
    • AI Lease Assistant chatbot answers manager and resident questions about any lease — rent amount, notice period, deposit, expiry date — without manual lookup
    • AI financial insights surface collection rate anomalies, payment pattern trends, and revenue forecasting to help managers make proactive portfolio decisions

    Why This Is the Best Solution

    Residenta is the only property management platform that combines AI ticket classification, predictive maintenance forecasting, automated daily cron jobs for rent and lease alerts, and a full resident self-service portal at a price point accessible to independent landlords and mid-market property companies alike. Enterprise tools like AppFolio and Buildium cost $1,000–$3,000/month and require implementation teams. Residenta is operational in minutes with a self-serve setup — and grows from 10 units to 10,000 without re-platforming.

    Smart Retail Engine logo
    Retail & E-commerceLive

    Smart Retail Engine

    Unified Commerce Platform with Computer Vision & AI

    The Problem

    Mid-market and enterprise retailers were operating with deeply fragmented commerce stacks — separate platforms for in-store POS, online storefront, inventory management, customer loyalty, and analytics, none of which communicated in real time. The result: inventory oversells across channels, manual stock reconciliation costing 10–15 staff hours per week, no unified customer profile across touchpoints, zero predictive replenishment, and high shrinkage rates because physical inventory counts relied on manual scanning cycles. Traditional retail analytics lagged 24–48 hours behind actual store events, making real-time operational decisions impossible. Promotions and markdowns were applied manually without demand signal analysis, leading to both stockouts on trending SKUs and deadstock on slow-moving lines simultaneously.

    What We Built

    Smart Retail Engine is a unified commerce platform that connects inventory, sales, customer data, and AI intelligence across every retail channel — physical, online, and marketplace — into one operational system. Computer vision handles physical shelf monitoring. AI handles demand forecasting, replenishment, and markdown optimisation. A single customer profile powers personalisation across every touchpoint. Retailers replace four or five disconnected tools with one coherent platform that operates in real time.

    Real-time multi-channel inventory sync20–35% shrinkage reductionAI replenishment 2–4 weeks ahead
    View Case Study How We Built It

    What We Solved

    • Built a unified inventory engine that syncs stock levels across physical stores, warehouses, and online channels in real time — eliminating oversells and manual reconciliation
    • Integrated computer vision shelf-scanning to detect out-of-stock positions and planogram compliance without manual counts, triggering automated replenishment alerts
    • Created a single customer identity layer that merges POS, e-commerce, and loyalty touchpoints into one profile with full purchase history and lifetime value tracking
    • Implemented AI-driven demand forecasting that predicts SKU-level sales velocity by location, season, and promotional calendar — auto-generating replenishment purchase orders
    • Delivered a real-time analytics dashboard surfacing sales velocity, margin by category, shrinkage anomalies, and footfall-to-conversion rates across all locations
    • Built an AI markdown optimisation engine that recommends time-based discount curves per SKU to clear deadstock at maximum recoverable margin

    How AI Transforms This Product

    • Computer vision shelf-scanning detects out-of-stock conditions and planogram violations automatically — no manual stock walks required, reducing shrinkage by 20–35%
    • AI demand forecasting models analyse sales velocity, seasonal patterns, promotional uplift, and local events to predict SKU-level replenishment needs 2–4 weeks ahead
    • AI markdown engine calculates the optimal discount curve per SKU to clear deadstock while maximising recovered margin — replacing gut-feel pricing decisions
    • Customer intelligence AI identifies high-value customer segments, churn risk signals, and next-best-product recommendations for personalised promotions across email, SMS, and in-store

    Why This Is the Best Solution

    Smart Retail Engine is the only unified commerce platform built with computer vision inventory monitoring and AI demand forecasting as first-class features — not third-party integrations bolted on from separate vendors. Competing solutions like Shopify Plus or Lightspeed handle transactions but require four to six additional tools for the intelligence layer. Smart Retail Engine delivers the full stack — POS, e-commerce sync, computer vision, demand forecasting, customer data platform, and markdown AI — under one multi-tenant, role-secured architecture designed to scale from a single-location boutique to a 200-store enterprise without re-platforming.

    Skill Gauge logo
    HR Tech & Talent ManagementLive

    Skill Gauge

    AI-Powered Talent Assessment & Skills Intelligence Platform

    The Problem

    HR teams and recruitment agencies were making hiring and promotion decisions based on CVs, self-reported skill ratings, and unstructured interviews — none of which reliably predicted on-the-job performance. Screening hundreds of applicants for technical roles required senior engineers to spend 2–4 hours per candidate on manual code reviews and test evaluation, pulling them away from product work. Internal talent mapping was non-existent: L&D teams had no objective view of which skills existed across the organisation, where gaps sat relative to roadmap requirements, or which employees were ready for promotion. Skill decay in fast-moving technology stacks was invisible until a project failed. Recruitment agencies were delivering candidates they had not objectively vetted, leading to early attrition and client churn.

    What We Built

    Skill Gauge is an AI-powered talent assessment and skills intelligence platform that replaces manual screening with adaptive, objective technical and behavioural assessments — auto-generated from job descriptions, auto-evaluated by AI, and benchmarked against role calibration data. For HR teams it provides an organisation-wide skills map, gap analysis against roadmap requirements, and promotion-readiness scoring. For agencies it delivers verified candidate profiles with objective skill scores — replacing unverified CV claims with assessed competency data.

    2-min AI assessment generation30–40% fewer questions, same accuracyOrg-wide skills gap map built in
    View Case Study How We Built It

    What We Solved

    • Built an AI assessment engine that generates role-specific technical and behavioural assessments from a job description in under 2 minutes — no manual question authoring required
    • Created adaptive question sequencing that adjusts difficulty in real time based on candidate performance, producing accurate skill scores from fewer questions
    • Implemented automated code submission evaluation with AI-powered review covering correctness, edge cases, code quality, and architectural reasoning
    • Delivered an organisation-wide skills matrix that maps assessed competencies across teams and departments, surfacing gaps against open roles and upcoming projects
    • Built a candidate benchmarking engine that scores applicants against a calibrated role benchmark and against the existing top-performer baseline for that function
    • Integrated anti-cheating detection: browser focus monitoring, copy-paste detection, IP consistency, and AI-generated response detection across written submissions

    How AI Transforms This Product

    • AI assessment generation creates role-specific technical, logical, and behavioural question sets from a job description in under 2 minutes — covering full-stack engineering, data science, product management, and 50+ other functions
    • Adaptive difficulty sequencing uses Item Response Theory models to adjust question complexity in real time — producing accurate skill percentile scores with 30–40% fewer questions than fixed-format tests
    • AI code review evaluates submissions across correctness, time/space complexity, edge case handling, code readability, and architectural decisions — providing detailed candidate feedback reports instantly
    • Skills gap intelligence AI cross-references assessed employee competencies against open role requirements and product roadmap timelines, prioritising L&D investment on the highest-impact gaps

    Why This Is the Best Solution

    Skill Gauge is the only talent assessment platform that combines AI-generated assessments, adaptive question sequencing, automated code evaluation, and organisation-wide skills intelligence in a single platform accessible to companies of all sizes. Enterprise ATS platforms like Workday or SAP SuccessFactors include basic skills taxonomies but no adaptive assessment engine. Dedicated testing tools like HackerRank or Codility cover technical coding tests but have no behavioural assessment, skills mapping, or internal mobility intelligence. Skill Gauge does all of this — and auto-generates assessments from job descriptions rather than requiring weeks of manual question authoring, making high-quality hiring science accessible from day one.

    GymApp logo
    Fitness & WellnessLive

    GymApp

    Multi-Tenant Fitness Management SaaS

    The Problem

    Fitness studios and gyms were operating with a fragmented toolset — member management in one system, class scheduling in another, and billing managed manually through disjointed payment gateways. Members faced clunky, outdated interfaces to book classes or view their workouts. For owners, there was no centralized way to track revenue, manage capacity, or handle multi-location operations without relying on legacy enterprise software that was overly complex, expensive, and difficult to customize.

    What We Built

    GymApp is a multi-tenant fitness and gym management SaaS platform that replaces fragmented tools with a single solution for class scheduling, QR code check-ins, automated billing, workout tracking, and member management. Designed to scale from single boutique studios to multi-location franchises.

    45% reduction in admin time20%+ increase in class bookings99% on-time payment collection
    View Case Study How We Built It

    What We Solved

    • Built a centralized dashboard for revenue, active members, and attendance tracking
    • Integrated automated Stripe billing to eliminate manual invoice chasing
    • Created a seamless class scheduling system with automated waitlists and capacity limits
    • Implemented dynamic QR code check-ins for frictionless front desk operations
    • Developed a comprehensive workout tracking module for custom programs and PR logging
    • Architected a secure 4-tier Role-Based Access Control (RBAC) system using Row-Level Security

    How AI Transforms This Product

    • Automated waitlist management intelligently predicts cancellations and automatically promotes members to fill classes
    • Analytics dashboard surfaces member attendance trends to predict and prevent churn before it happens
    • Smart capacity planning recommends optimal class schedules based on historical attendance data and facility size

    Why This Is the Best Solution

    GymApp provides a streamlined, highly performant experience with modern architecture and zero bloat. Unlike legacy platforms, it offers deep customization options, full white-label capabilities (custom domain, logo, colors), and the option for a dedicated, self-hosted deployment tailored entirely to your brand.

    Active Research Areas

    The technologies our next products will be built on. Research clients get early access before general availability.

    Agentic AI Systems

    Active Research

    Multi-agent orchestration patterns for enterprise automation. We are evaluating LangGraph, CrewAI, and AutoGen at production scale — latency, cost, and reliability benchmarks across real workloads.

    Expected Outputs

    Benchmark report (2026 Q1)Open-source agent evaluation harnessProduction deployment patterns guide

    Edge AI & On-Device Inference

    Prototyping

    Running quantized LLMs (Phi-3, Llama 3 4-bit) on edge devices for offline-capable AI applications. Use cases: field service, healthcare diagnostics, industrial IoT.

    Expected Outputs

    Latency benchmarks by device classQuantisation impact reportOffline-first AI architecture guide

    Spatial Computing & AR

    Exploration

    Apple Vision Pro and Meta Quest 3 development for enterprise use cases: remote assistance, training simulations, and collaborative data visualisation.

    Expected Outputs

    Vision Pro enterprise app prototypeSpatial UI/UX design patternsEnterprise AR deployment guide

    Developer Tooling

    Active Development

    Internal tools for AI-assisted code review, automated architecture documentation, and test generation. Some will be open-sourced.

    Expected Outputs

    AI code review CLI toolArchitecture diagram generatorLLM test generation library

    Also From Our Lab

    An earlier generation product now serving clients in digital health and clinical documentation.

    Ortem Health logo

    Ortem Health

    AI-assisted clinical documentation and patient engagement platform. Built on our HIPAA-compliant AI research. Serving digital health providers, clinics, and telehealth platforms.

    Learn more →
    See All Products

    Want a Product Like This Built for Your Business?

    Fleet Sync Pro, Azimuth, Residenta, Smart Retail Engine, and Skill Gauge prove what our team builds when given a real problem and full creative control. The same team is available for your project.

    Production-Grade from Day One

    Every product we build ships with multi-tenant RLS, RBAC, automated testing, and cloud-native infrastructure — not bolted on later.

    AI-First Architecture

    We do not add AI as a feature. We design systems where AI handles classification, prediction, generation, and automation at the core.

    13 Years, 1,000+ Projects

    Our Lab products are built by the same team that has shipped software for clients across the USA, UK, Australia, Canada, and Middle East since 2012.

    Related: AI & ML Services · AI Agent Development · SaaS Development · All Case Studies · MVP Development

    Work With Our Innovation Team

    If you are working on a product in fleet management, agency tooling, property technology, retail commerce, talent assessment, AI automation, or any domain that touches our research areas — we would love to explore a collaboration. Research clients sometimes get access to technologies before they are generally available.