Services

One flywheel: assess, automate, train. Built to be funded.

Every service ships with a published scope, price band, and timeline, sequenced around one federal-alignment spine. Assess carries zero code and zero liability and is structured to support grant applications. Automate stays bounded and human-in-the-loop, engineered to qualify for the Productivity Super-Deduction. Train leaves your team fluent on subsidizable workshops. Sovereign, pro-worker, deterministic — built to be used, owned, and funded.

On this page 0x01 Assess 0x02 Automate 0x03 Train 0x04 Assist 0x05 Index 0x06 Foundations 0x07 Rapid Web 0x08 Enable +ADD AIDA Audit
0x01 / Assess · Featured flywheel

Grant-Ready Roadmap

The on-ramp that feeds everything. A one-week engagement that scores candidate workflows on payoff versus effort, then structures the ROI projections and risk-mitigation outlines to support BDC, cluster, and grant applications. Zero code, zero liability.

Funded by · BDC LIFT / Global Innovation Cluster & SME grant applications
Length
1 week, four working sessions
Where
On-site at your office (GTA) or remote
Investment
From 5,500 CAD
Deliverable
Written plan with three to five scored opportunities, ROI projections, and risk-mitigation outlines structured to support grant and cluster applications
Yours to keep
The written plan is yours to act on — with us, another builder, or no one. Digital infrastructure, not vendor dependency. No follow-on required.
Best for
Teams who’ve heard the AI pitch from a dozen vendors and want a grounded, grant-ready second opinion

What happens in the week

  1. Day 1. Two hours with leadership. We map the org and the workflows that hurt.
  2. Day 2 to 3. Shadow the teams that own those workflows. Read the docs. Watch the work.
  3. Day 4. Score each opportunity on payoff, effort, risk, and adoption likelihood — and map each to the funding that could pay for it.
  4. Day 5. Present the written plan and walk the room through it.

Book a Grant-Ready Roadmap


0x02 / Automate · Execution centerpiece

Bounded Workflow Automation

The boring half of every operations team’s calendar moved into a deterministic workflow that runs without them. n8n, Make, or custom code stitching email, docs, CRM, and Slack — with LLM steps only where they earn it. Level 1–2 assistive, human-in-the-loop by design.

Written off via · Productivity Super-Deduction
n8nMakeZapierCustom Node / PythonWebhooksPostgres triggers
Length
2 to 4 weeks
Investment
12 to 26k CAD, fixed
Includes
Build, monitoring, run-book, 30 days of support — deterministic and easily-run, with the keys handed to your team
Autonomy posture
Level 1–2 assistive, read-only-by-default, human approval gate before anything is written or sent
Common shapes
CRM hygiene, inbound triage, document routing, recurring report generation, customer onboarding handoffs
What we won’t do
Replace a workflow that’s working fine on a spreadsheet

Scope a bounded workflow


0x03 / Train · Featured flywheel

Defensive AI Literacy & Upskilling

Pro-worker, hands-on workshops for the people doing the work, not the people writing the policy. We bring a laptop, your team brings their actual tasks, and they leave with FASTER principles, defensive prompting, and PIPEDA-safe habits they can use on Monday.

Subsidized by · Canada-Ontario Job Grant (COJG)
Format
Half-day (3 hours) or full-day (6 hours), on site in the GTA
Investment
3,500 CAD (half)  ·  6,000 CAD (full)  ·  9,000 CAD (custom cohort)
Funding
May be eligible for up to ~83% of training cost via the Canada-Ontario Job Grant (COJG), subject to current program terms
Audience
Operations, sales, legal, finance. Not engineers. Aimed at the curious non-technical professional.
Group size
5 to 20 people. We run two sessions for larger teams.
What you take home
A short written field guide, three workflow templates, and a recording of the session
Curriculum (full-day)
  1. What these tools actually do, in 20 minutes
  2. Prompting workshop with your own real tasks
  3. Where AI fails, and how to tell from a distance — defensive prompting and treating external text as data, not instructions
  4. Building three bounded workflows you’ll use next week
  5. Privacy, security, and what to do with client data — PIPEDA-safe by habit

Book a workshop ↗


0x04 / Assist

Bounded Assistants & Internal Q&A

Customer-facing chat and internal Q&A over your documents, built bounded and recoverable — not black-box autonomy. Level 1–2 assistive, read-only-by-default, with a human approval gate before anything is written or sent. A real assistant in production, shipped with evals and monitoring.

Aligned with · the federal Guide on the Use of Agentic AI
Claude Sonnet 4.5GPT-5LangGraphVercel AI SDKpgvectorEval harnessLangFuse
Length
4 to 8 weeks
Investment
28 to 72k CAD, fixed
Includes
Evals, monitoring, fallback strategy, audit logs, kill switch, handoff doc, 30 days of support
Autonomy posture
Level 1–2 assistive only. The assistant analyzes, retrieves, and drafts; it is architecturally blocked from altering databases or sending external messages without a human key. Never Level 3/4 or black-box.
Two bounded shapes
Concierge — customer-facing chat that drafts and routes, with a human approval gate on anything sent
Reference — internal Q&A over your docs, contracts, and tickets, read-only with cited sources
What you get on day 21
A working pass shown to a small pilot group, end-to-end

Scope a bounded assistant


0x05 / Index

Internal Knowledge Base & Company OS

One secure, permission-aware search layer over everything your company knows — the handbooks, contracts, tickets, scanned PDFs, and line-of-business systems your team currently digs through by hand. Ask in plain language, get an answer with its source, and never see a document you couldn’t already open.

Built for · Canadian data residency & access-control inheritance
pgvector / SupabaseQdrantAzure AI SearchRerankingDocument-level ACLClaude Sonnet 4.5Eval harness
Length
4 to 8 weeks to a first production index
Investment
15 to 40k+ CAD build, plus a managed retainer
The hard part we own
Permission inheritance — the index honours your access rules at query time, so it can’t surface a file a user couldn’t already open. Get this wrong and you’ve built a data-leak machine; getting it right is the whole job.
Where we fit
If your docs are tidy and in one system, an off-the-shelf add-on may be all you need — we’ll tell you so. We earn our keep where data is messy, spread across systems, and permission-sensitive: scanned files, legacy stores, and line-of-business apps generic tools can’t reach.
Architecture
Ingestion and chunking, embeddings, a sovereign-hosted vector store, retrieval with reranking, and document-level access-control filtering on every query. Read-only, cites its sources, yours in your cloud.
The retainer
Re-indexing as your documents change, new sources connected, and retrieval quality tuned against the questions people actually ask — so the index stays useful instead of going stale.

Scope a knowledge base


0x06 / Foundations

AI-Ready Digital Foundations

Marketing sites and product surfaces, sovereign-hosted and built to target WCAG 2.1 AA, architected to integrate bounded automations later instead of bolting them on. Low regulatory risk — the front end of an AI system done right from the first commit.

Architected for · Canadian data residency & WCAG 2.1 AA
Next.jsAstroSanity / PayloadTailwindAWS Canada CentralVercel / Cloudflare
Length
3 to 6 weeks
Investment
18 to 55k CAD, fixed
Includes
Discovery, design system, build, CMS, sovereign deployment, analytics, training
Accessibility
Semantic HTML and ARIA labels, built to target WCAG 2.1 AA
When to pair with assistants
Bounded concierge chat, internal Q&A, semantic search, lead qualification — added later without a rebuild
What we won’t do
Themed WordPress builds. Generic SaaS landing pages. Templated work that doesn’t need us.

Discuss a digital foundation ↗


0x07 / Foundations · Rapid track

Rapid Web Build

An AI-ready digital foundation in two to three weeks — sovereign-hosted, built to target WCAG 2.1 AA, and yours to own in your repo on your domain. Real design, real copy, real code. The savings come from how we build, not from cutting corners on what you get.

We use AI internally to compress the slow parts — first-draft layouts, copy variations, component scaffolding, animation prototypes — then a senior practitioner edits, directs, and ships. No bloat, no six-month discovery, no hand-off to juniors after the sales call. You get the output, not the overhead.

Architected for · Canadian data residency & WCAG 2.1 AA
Next.jsAstroTailwindFramer MotionCMS-optionalAWS Canada CentralVercel / Cloudflare
Length
2 to 3 weeks, end to end
Investment
6,500 to 18,000 CAD, fixed
Includes
Brand-aligned design system, hand-written copy, responsive build, basic SEO, sovereign deployment, 30 days of changes
Best for
SMBs, professional services, and operators who need a real website without a four-month agency cycle
What we won’t do
Themed templates. Page builders. Anything that locks you into our hosting. The site is yours, in your repo, on your domain.
Difference from 0x04
This is a clean, AI-ready site built fast. 0x04 is the same foundation with a bounded AI system inside it (concierge chat, internal Q&A, semantic search). Different scopes, different prices.
How the two-week build runs
  1. Day 1–3. Strategy call, brand audit, sitemap, copy outline.
  2. Day 4–7. First full pass — design + copy + working pages. Review.
  3. Day 8–12. Second pass, polish, mobile, edge cases, copy edits.
  4. Day 13–14. Deployment, analytics, handoff, content training.

Quote a website


0x08 / Enable

AI Enablement

The work that decides whether a system becomes the way work gets done or sits idle next to a Slack channel of complaints. A six to twelve week engagement focused on adoption, not implementation — so the day we leave looks like the day after, for the people doing the work.

Length
6 to 12 weeks, retainer
Investment
From 9,500 CAD / month
Who it’s for
Companies that already built something with AI but no one is using it (and you suspect why)
What we do
Adoption interviews, rollout playbook, training cadence, weekly office hours, metrics dashboard
Outcome
A measurable shift in daily active use, not slide-deck testimonials

Three failure modes we fix

  • The pilot purgatory. Something works on day one and nobody touches it on day thirty. We find the friction and remove it.
  • The leadership-vs-line gap. Executives think it’s adopted; the people doing the work disagree. We close the loop with measurement.
  • The trust deficit. The model is correct but the team doesn’t believe it. We rebuild trust through visibility, the approval gate, and override controls.

Start an enablement engagement


+ADD / Governance add-on

AIDA Compliance Audit Add-on

A modular governance add-on for teams deploying AI under scrutiny. We screen for algorithmic impact, red-team the system against prompt injection and automation drift, and design the audit logs and kill switch that make a deployment recoverable. Layer it onto any build, or take it standalone on the compliance track.

Built for · AIDA readiness & the Voluntary Code of Conduct
Format
Modular — add to any build, or take standalone
Investment
8,000 to 18,000 CAD, scoped to system size
Scope
AIA screening — an Algorithmic Impact Assessment to place the system on impact Levels I–IV
Red-teaming — adversarial testing for prompt injection and automation drift
Audit logs — immutable, human-readable records the AI cannot alter
Kill-switch design — a literal control to sever AI access and revert to manual
Data-residency review — where data lives and moves, mapped to PIPEDA and Quebec Law 25
Maps to
The Voluntary Code of Conduct (human oversight, robustness/red-teaming) and AIDA readiness
Best for
C-suite, IT, and legal teams deploying AI in a regulated or data-sensitive setting

See the AIDA Compliance Audit


How they combine

Three common paths.

Most clients chain two or three engagements over six to nine months. The flywheel runs the same way each time — assess, automate, train — paid for by funding or written off as productivity spend. These are the three sequences we see most often.

Path A

Growth track, funding first

Grant-Ready Roadmap → Defensive Literacy & Upskilling → Bounded Workflow Automation
~25k CAD  ·  6 weeks  ·  grant-ready, subsidizable, written off
Path B

Compliance track, deploy securely

Grant-Ready Roadmap → Bounded Assistant + AIDA Compliance Audit → Enablement retainer
~85k CAD  ·  16 weeks  ·  bounded, recoverable, AIDA-ready
Path C

Building product, AI inside

AI-Ready Digital Foundations + Bounded Assistant, paired
~60k CAD  ·  8 weeks  ·  sovereign-hosted, ship-to-prod

Pick a track

Two ways in.

Same flywheel, same practice behind both. Pick the one that sounds like your week — whether the pressure is growth and funding or risk and compliance.

Not sure where to start?

Start with the Grant-Ready Roadmap. Most clients do.

One week, fixed price, an actual plan you can read — with each opportunity mapped to the funding that could pay for it. If the roadmap finds that AI isn’t the right answer for you right now, we’ll say so. We get paid for the week either way, and you don’t waste a year.

Map Your AI Funding & Compliance Strategy See pricing table