Our work.

YoungStars runs two childcare centres, one in Vancouver, one in Kelowna, serving kids from infancy through before-and-after school care. When we first spoke, the director was managing enrolments for four programs across both locations using a mix of paper forms, spreadsheets, and email threads. Nothing talked to anything else.

The immediate problem wasn't the paperwork itself. It was that the director had no clear picture of what was happening across both sites at once. Inquiries came in through different channels. Staff schedules lived in a shared Google Sheet that was always slightly out of date. There was no reliable way to know at a glance how many spots were open, which families were in the intake pipeline, or what the wait situation looked like.

We built a single platform that covers both locations. Parents can register online — medical info, dietary needs, emergency contacts, all of it — and the system routes everything to the right program and site automatically. Staff schedules are managed in the same place, with shift types and hourly rates handled per role. The director gets a dashboard that shows both locations together, not as two separate things to check.

We also added an AI-powered inquiry intake that reads the message a parent sends and extracts their desired start date and program interest before it ever reaches a human. That alone saves the team a meaningful amount of back-and-forth every week. A testimonials engine and location-tagged blog handle the SEO side quietly in the background.

The platform replaced what would have taken four or five separate tools — and none of those tools would have talked to each other.

GS Property manages a student housing portfolio in Kelowna: landlords on one side, tenants on the other, with all the leases, maintenance requests, finances, and communications that sit between them. Before we worked together, most of that was running on email, manual documents, and a collection of spreadsheets that required constant upkeep.

The lease process alone was a significant time drain. BC's RTB-1 form needs to be completed, signed, and stored for every tenancy. They were doing this manually, chasing signatures, and then filing paper copies. On top of that, each landlord needed monthly invoices with GST calculated correctly, and there was no clean way to generate those without doing it by hand each time.

We built a full property management platform. Landlords and tenants each have their own portal. Prospective tenants can browse listings and book viewings directly from the site — or through Facebook Messenger, which is where a lot of their inquiries were already coming in. The RTB-1 lease signing is now digital: a unique token gets sent to the tenant, they sign online, and the signed document is stored automatically. The 48-hour expiry on tokens was Tariq's idea — it creates the right amount of urgency without being aggressive.

The financial side generates invoices, credit memos, and landlord statements automatically, with GST applied correctly every time. Maintenance requests go into a ticketing system. Student co-tenants are managed as roommate groups under a single lease. And the AI chatbot — which runs on a versioned prompt system so responses can be tuned without a developer — handles the first layer of tenant inquiries around the clock.

It replaced what would have otherwise been five or six separate tools, none of which would have been built for property management in BC specifically.

MindGuard had developed a four-module course called "AI for the Critical Thinker", designed to teach high school students how to use AI responsibly rather than passively. The content was aligned to Ontario Ministry of Education guidelines and supported by the Canada Service Corps and Frontline Community Centre. The philosophy was clear: AI is a tool to augment human intelligence, not replace it.

The problem was delivery. The curriculum was being run in person, in limited settings, with no way to reach the schools and community centres at the scale they needed. On top of that, funders wanted proof — who completed the course, what scores looked like, what the demographic breakdown was. That's hard to answer when everything happens in a room.

We built a full e-learning platform on Base44. Four modules — "The Hidden Engine," "The Hallucination Hunter," "The Socratic Tutor," and "The University Bridge" — each with video lessons and multiple-choice assessments that need to be passed before the next module unlocks. The platform tracks individual student progress through every segment and quiz, with completion percentages visible on a personal dashboard.

New students go through a profile setup and demographic survey on first login, which feeds directly into the admin reporting. When someone completes all four modules, they get a personalized certificate of completion with a unique ID and their score — stored as a permanent record.

The admin dashboard gives program coordinators a full picture: enrolment stats, per-student progress tracking, issued certificates, and aggregated survey responses broken down by student status, institution type, and referral source. The whole thing is architected so a school administrator or community centre coordinator can run it without needing technical support.

The curriculum that existed in a room now exists on the internet, with proof of who completed it and the data funders need to see impact.

STROM Futures was building ACTx02, a platform targeting Canadian oil and gas SMBs around carbon emissions strategy. Good product, right timing, but no validated process for getting into rooms. No discovery framework, no target list, no outreach system that could work at scale with a small team.

The engagement ran five weeks through Riipen. The deliverables were specific: a sales discovery framework, market intelligence on where O&G SMBs were actually failing, and an outreach system that could produce personalised messages without requiring a salesperson to research every company from scratch.

The discovery framework ended up at 30 questions built across three methodologies (SPIN, Sandler, and MEDDPICC) but written for the oil and gas context specifically. Not adapted from generic B2B templates. The difference matters when you're in a room with an operations manager at a mid-size producer and you need to ask about their actual pain, not a category approximation of it.

The market intelligence piece identified five failure modes common across O&G SMBs, pulled from over 50 validated sources, and produced a ranked target list with contact intelligence ready to use. The research pipeline ran on ScrapeGraphAI — scanning company sites, job boards, and industry directories for signals that a company was in the right situation for the conversation.

The outreach system connected Happenstance with Claude. It produced personalised messages at scale, grounded in real company signals rather than mail-merged templates. The whole thing was documented so the team could run it themselves after the engagement ended.

What Tariq said afterward: the work consistently went beyond the stated scope, not in a way that added noise, but in a way that extended the original direction rather than just executing it. That calibration is what we try to bring to every engagement.