AYODEJI
OGUNLAMI

SurveyEase.ai is a web and mobile SaaS product that gives land surveyors one AI-assisted workspace for producing regulator-ready plans. Surveyors can upload a hand-drawn field sketch, scanned drawing, or raw coordinate list and receive a cleaned, scaled plan instead of redrawing the material manually in CAD. The same workflow checks the plan against a defined regulatory ruleset and provides an assistant grounded in the relevant survey regulations. Together, these tools move plan preparation and early compliance review from hours of manual work to a process that can be completed in minutes.

• Image-to-Plan converts phone photographs and scanned drawings into cleaned, scaled survey plans that can be exported to PDF or CAD
• Coordinates-to-Plan creates filing-ready plans with the correct bearings directly from raw coordinate lists
• The compliance checker scores each plan against a 25-point regulatory ruleset and highlights issues before the surveyor submits it
• The regulatory copilot answers practical questions and cites the relevant survey and land legislation in each response
• The product runs as a Next.js web app and an Expo mobile app, with sign-in, local-currency billing, team seats, and data-protection controls

MindEase Pro is a full-stack mental well-being platform that gives users several ways to reflect and talk: text, low-latency voice, or a live video-avatar conversation. The product is designed for moments when someone wants immediate, private support but cannot easily reach affordable professional support. Alongside the conversational experience, it includes recognised screening questionnaires, crisis checks, mood and journal tools, and real-time couples sessions. Trauma-aware interaction patterns, accessibility, privacy, and clear escalation routes are considered throughout the interface and backend rather than being added as separate features at the end.

• Supports conversations by text, low-latency voice, or a live session with a talking video avatar
• Couples mode allows two partners to join from separate phones while an AI moderator guides turn-taking and shared exercises in real time
• Includes recognised PHQ-9 and GAD-7 screening, checks for signs of self-harm, and routes users to crisis lines based on their region
• Offers interactive breathing, grounding, CBT thought-work, and couples-card exercises through a cache-first content engine
• Maintains memory across sessions alongside mood tracking, journaling, sign-in, and billing, with an AI gateway selecting models according to cost and urgency

Oracle is a personalised self-reflection product that uses a person's name and date of birth to create detailed readings covering daily guidance, annual forecasts, compatibility, career direction, and support during difficult periods. It is designed for people who already use products such as Co-Star or The Pattern as part of a regular reflective routine. The experience keeps each reading private, immediate, and consistent in tone across every channel. What started as a live Telegram bot now serves Telegram, web, and mobile from one shared reading engine.

• Turns a name and date of birth into nine reading types, including daily guidance, compatibility, career direction, annual forecasts, and support through difficult periods
• Runs as a Telegram bot, a web application, and native iOS and Android applications, all maintained from one shared codebase
• Uses one reading engine with more than 50 chart aspects, with automated parity tests confirming that every platform produces the same result
• Brings card payments, mobile in-app purchases, and Telegram's own currency into one subscription and entitlement system
• Generates readings on the user's device and anchors each daily reading to the user's actual local date rather than the server's timezone

FIFA Arena is a competitive platform where online EA Sports FC players can build a record and compete for cash prizes through structured league seasons, a class-based challenge ladder, and skill-matched contests. Many communities currently manage rankings, match results, disputes, and prize payments across separate WhatsApp or Discord groups. FIFA Arena brings those activities into one system with clear rules, confirmed results, player ratings, and controlled settlement. A shared TypeScript backend powers the working web application, while a mobile client remains part of the product roadmap.

• Runs a four-tier challenge ladder with Glicko-2 skill ratings, promotion rules, and checks for possible collusion between players
• Manages complete league seasons, including divisions, generated fixtures, live standings, and defined tie-break rules
• Records entry fees and prize settlements on a double-entry ledger designed to keep every movement of player funds traceable
• Includes phone-based sign-in, an 18+ gate, mutual result confirmation, a dispute process, and an administrative resolution queue
• Reshapes informal head-to-head wagers as controlled prize contests, with licensing and custody requirements separated from the core competition product

EstateIQ is an AI business-intelligence and workflow product for real-estate companies. It turns a large property database into information that sales, finance, and operations teams can explore by asking questions in ordinary English. For example, a manager can ask which zones have the highest outstanding balances and receive a written explanation, the supporting rows, and a chart without waiting for an analyst to prepare a report. The product brings parcel inventory, customer payments, title status, construction progress, and infrastructure faults into one consistent data model that the wider company can query.

• The ask-your-data chat converts plain English into a validated, read-only database query, then returns a written answer, a table, and a chart
• Query guardrails restrict access to approved tables, block statements that could change data, enforce row limits, and reduce unnecessary data exposure
• Provides a KPI dashboard, parcel inventory, a collections-priority queue, and a workflow for managing facilities and reported faults
• Combines estate map layers and operational spreadsheets in one queryable database with consistent fields and relationships
• Retains conversation history for follow-up questions, masks buyers' personal information, and remains demonstrable without an AI key through a local fallback

Zuka is a full-stack, multi-brand fashion marketplace that brings independent clothing labels together in one curated storefront. The platform manages product discovery, checkout, payments, fulfilment, and post-purchase updates across participating brands. It gives shoppers one polished place to browse emerging labels, while giving smaller brands access to a modern commerce workflow without requiring each one to build and operate its own platform. Alongside the conventional storefront, Zuka includes an AI shopping companion and a game-based loyalty programme designed to encourage discovery, repeat visits, and longer-term customer relationships.

• Provides an AI shopping assistant with selectable personas, memory for each shopper, and layered checks against prompt manipulation
• Supports the complete shopping journey through filtered search, a multi-step checkout, card and local payments, coupons, wishlists, and order tracking
• Includes administrative and brand-onboarding portals for bulk product uploads, order management, analytics, and customer-review moderation
• Runs a game-based loyalty system with playable arcade titles, points, tiered ranks, leaderboards, and reward redemption
• Uses one monorepo for the Next.js web application and React Native mobile application, supported by more than 550 tests and a complete CI/CD pipeline

Vyra is an Afrobeats music-creation studio for independent artists who want to move from an early idea to a finished master in one product. An artist can generate a complete track from a written prompt or record an original vocal and use the studio to produce, mix, and master it. The workflow continues beyond audio generation: each export includes a distribution-ready package with AI-disclosure and licensing information prepared for the artist. This brings creation, recording, post-production, documentation, and export together instead of leaving the final packaging work to separate tools.

• Offers two creation paths: generate a complete track from a text prompt, or record a vocal and have the studio produce and master it
• Packages every export as a distribution-ready master with the relevant AI-disclosure and licensing records prepared alongside it
• Runs across iOS, Android, and web on one shared backend with a typed API and consistent project data
• Places several music and voice AI providers behind a swappable adapter, with clear error states and recovery when a provider fails
• Supports African markets through live local-currency checkout, tiered subscriptions, credit top-ups, and genre presets developed for regional styles

Hermes organises a broad set of operational capabilities inside one self-hosted AI assistant: 14 departments, 31 specialist roles, and 142 ready-to-run skills covering marketing, engineering, legal, finance, sales, design, and media. It is intended for founders and small teams that want a private assistant capable of handling structured work across several business functions. When a request arrives, Hermes identifies the relevant department and role, loads the matching instructions, and follows that playbook from the initial brief through to a reviewable deliverable.

• Includes 142 ready-to-run skills across 14 business functions, with each skill providing a complete and reusable working playbook
• Uses a company charter as an organisational map, routing each request to the appropriate department and specialist role before work begins
• Applies permanent approval limits to spending, external messages, and signatures, while requiring sources for factual claims and sensitive recommendations
• Uses a tool-independent connector layer that falls back to web research and local files when an external service is unavailable or not connected
• Provides a custom packaging toolchain that bundles the departments, roles, and skills into a deployable installation with licence and integrity checks

AI Recruiting Automation manages the early stages of a hiring process from one connected workflow. A candidate applies through a role-specific web form, the screening pipeline reads and scores the CV against the published requirements, and shortlisted candidates are invited to a short AI voice interview that runs directly in the browser. Recruiters receive the score, written reasoning, transcript, and interview assessment in one control panel. The platform reduces manual CV triage and first-round scheduling while keeping every progression, rejection, and final hiring decision with a human recruiter.

• Reads each CV, scores it out of 10 against the requirements of the specific role, and records a written rationale for the recruiter
• Runs an approximately 10-minute conversational voice interview in the browser, then transcribes the discussion and scores the agreed competencies
• Uses a custom white-label interview interface with no third-party branding and server-side personalisation that candidates cannot alter in the browser
• Turns an Airtable board into the recruiter's control panel; moving a candidate card triggers the correct branded invitation or outcome email once
• Uses n8n to coordinate the AI, voice, database, and email services, with separate models for high-volume screening and deeper interview assessment

Regulatory Change Navigator is a multi-agent compliance-monitoring system for insurance and financial-services teams that must follow a continuous stream of publications from several regulators. It monitors official feeds, separates new or changed material from previously reviewed items, and explains each development in plain language. The system then compares the new requirement with internal policies, identifies possible gaps, and prepares draft amendments for a compliance professional to review. One dashboard brings together monitoring, analysis, policy comparison, proposed wording, and the evidence behind each finding.

• Monitors five regulatory feeds across the UK and US, deduplicating the results and retaining only material that is new or has changed
• Uses a six-agent workflow to summarise each rule, assess its likely impact, and identify potential gaps in the organisation's internal policies
• Prepares redlined policy amendments and sends Microsoft Teams or email alerts when a finding is rated high or critical
• Provides a complete web application with live pipeline progress, gap review, a policy-difference viewer, and a plain-English query panel
• Runs on Databricks with open foundation models, vector search, structured application data, and MLflow tracing across the complete workflow

Speak to PDF is a voice-enabled document assistant for legal professionals. A user can upload contracts or other legal material, listen to an automatically generated summary, and ask follow-up questions by voice or text. The assistant retrieves supporting passages from the uploaded documents before producing an answer, so the user can trace the response back to its source. It also compares relevant answers with the UK Data Protection Act 2018 and presents the associated requirements and citations, turning a lengthy document-review task into a focused, evidence-backed conversation.

• Accepts legal questions by voice or text and supports each response with cited passages retrieved from the uploaded document
• Checks relevant answers against the UK Data Protection Act and presents the applicable requirements alongside the document-based response
• Processes several PDFs in one session, creates a summary for each document, and answers questions across a user-selected collection
• Includes a custom in-browser voice component for hands-free recording, transcription, audio playback, and continued spoken conversation
• Uses Azure OpenAI and Azure AI Search for retrieval and response generation, with cached answers reducing repeated cost and latency

AI Employee Marketplace represents a virtual company as installable software: each department is a plugin, each specialist role is an AI agent, and each repeatable capability is packaged as a reusable skill. It is designed for Claude Code users who need structured support across engineering, design, marketing, media, sales, finance, legal, and operations. Instead of maintaining a loose collection of one-off prompts, a user can install an organised department with defined roles, tools, and working methods, or assemble the complete company one department at a time.

• Provides 14 business departments, approximately 30 specialist AI agents, and more than 140 skills that users can invoke directly
• Supports a one-command installation for either the complete virtual company or one selected department and its dependencies
• Keeps model selection configurable: agents inherit the active session, while generation models are chosen once and remembered for later work
• Stores skills in an open and portable format that can be used by other AI systems supporting the same specification
• Includes governance through tracked licence provenance, approval-gated external tools, validated package manifests, and a pre-commit integrity audit