GrantIQ AI grant matching and writing for Canadian artists

Canadian grants are spread across federal, provincial, municipal and private funders. Each has its own site, deadlines, eligibility rules and tone. Artists find them late, miss the fit, and rewrite the same artist statement and budget from scratch for every application.

The existing tools are databases. They list grants, but they don’t tell an artist which ones they can actually win, and they don’t help with the writing, which is where most of the time goes.

• Product strategy: positioning, niche, pricing tiers and roadmap.
• UX and UI: flows, wireframes, high-fidelity screens and the design system in Figma.
• Build: Next.js App Router, strict TypeScript, the data model, the scraping pipeline, AI features and payments.
• Delivery: 26 scoped sprints (19 core, 7 for visual arts), each starting with a structured Q&A before any code.

A review of Canadian grant tools and conversations with working artists surfaced three insights that shaped everything.

• Competitors stop at discovery. Grant tools in Canada are search-and-filter databases. The gap was matching and writing, not listing.
• The profile is the product. Match quality and writing quality both depend on how much the system knows about the artist. A rich profile, entered once, makes everything else work.
• “Artist” is too broad. A painter, a ceramicist and a media artist apply to different streams, use different vocabulary and are judged differently. Generic advice reads as generic to juries.

I launched for one discipline instead of “all grant seekers.” Depth beats breadth early: narrowing let me build discipline-specific vocabulary, funder tone profiles and relevance signals that a general tool can’t match.

Canadian-first sourcing, from federal, provincial and .ca funders, became a differentiator rather than a limit. Onboarding branches by discipline, so the profile and the grant feed are relevant from the first session.

The artist profile holds the statement, CV, exhibitions, residencies, budget and portfolio. It’s the central object: it feeds matching, feeds the AI writer and publishes a shareable portfolio page.

Artists used to retype the same material for every application. Now they enter it once, and the AI gets the context it needs to write something useful.

Every grant carries a match score built from eligibility, discipline relevance and profile signals. A long list of grants creates work. A ranked, explained shortlist tells artists where to spend their limited time.

Deadlines are surfaced by urgency, so nothing is found too late, and a Kanban pipeline tracks every application from saved to submitted.

The AI Writer drafts an application one section at a time, from the project narrative to the budget justification, using the artist’s profile, the funder’s tone and discipline-specific vocabulary. Drafts stream in, so it feels collaborative and fast.

The artist edits; the AI never submits. Juries respond to specificity and fit, so a draft grounded in the artist’s real work and the funder’s language is a starting point worth editing, not boilerplate to rewrite.

• Guidance beside every section: what the section needs, how to structure it, and the common mistakes to avoid.
• Tone and length controls: formal, conversational or academic, and a target word count for each section.
• Nothing is final until the artist says so: every draft keeps its history and can be regenerated or revised.

A compact, token-based system designed in Figma and built in Tailwind, shared by the artist app and the funder portal.

• Colour. A vermilion accent (#E8411A) for action and urgency, a near-black sidebar for focus, and a warm paper canvas (#F5F2ED).
• Type. DM Sans for interface and reading; DM Mono for amounts, deadlines and match scores, so numbers scan and align.
• Components. Grant cards, match-score indicators, deadline chips, Kanban columns, profile entry blocks and a streaming editor.
• Workflow. Figma, then HTML prototypes, then Claude Code. Design edits land as targeted section changes, so tokens and spacing never drift.

Profile completeness turned out to be the metric that matters: every feature downstream is only as good as what the artist gives it. Designing and building alone kept the loop short, since a decision made in Figma was tested in code the same week.

• The full artist platform: dashboard, discovery with filters, match scoring, the application pipeline and the AI Writer.
• Stripe billing across four tiers.
• A two-sided marketplace: a funder portal (funders.grantiq.ca) for posting grants, reviewing applications, analytics and verified badges.
• A community layer: an artist directory, an event calendar and discussion threads on individual grants.

Narrowing the audience improved the product more than any feature did.

• Track every step from match to saved, drafted, submitted and funded, and use it to tune the scoring.
• Test the AI Writer with artists to find where they edit most, then improve those sections.
• Expand to other disciplines with the same discipline-profile pattern.
• Launch ScholarIQ, a scholarship product for students that reuses the scraper, matching, AI Writer and pipeline.