Opening sequence
- Trigger
- Page load, once
- Response
- Headline, promises and mascot settle; the tour frame rises behind it
- Why
- A calm entrance for a buyer who is already anxious about security
A cybersecurity consultant with 15 years in the Canadian Army, government and police wanted to serve small businesses the way he served ministries. We researched the market, designed the product ladder, and built the website, free assessment, customer portal and review console that make one expert scale.
Manual security assessments cost $15K to $25K and take weeks. Compliance software costs $7.5K a year and comes with no human. Canadian SMBs got neither. Jirehnexus had to be the third option: an automated assessment, a named expert's signature, and a price three to six times lower.
A buyer who pays less expects less. The product had to feel more rigorous than a $20K consultant while costing a fraction, without a single fabricated testimonial or unsourced statistic.
Reduce 153 safeguards to the 30 that need Ruphin. A free 12-question assessment as the front door, automation for scoring and drafting, and a human approval that is a stored fact, not a convention.
A 24-page marketing site, six free tools, a customer portal and a review console, on one design system, with every open commercial decision logged for the founder.
A founder meeting, live market research, a strategy document as the single source of truth, then design and build in milestones with a self-review after each one.
Founder meeting and research.
The product, in writing.
Four surfaces, one map.
Find the look.
No framework, no build step.
Prove it before trusting it.
Why now, who pays, what they pay for, and the twelve decisions that turned a consultant's hours into a product.
Canadian small businesses are attacked like enterprises and insured like enterprises, but the only security help on offer is priced for enterprises.
Every figure on the client's site carries a citation the founder approved. Unsourced breach-cost numbers on the old site were removed. On a site selling trustworthiness, a wrong number costs more than a missing one.
We mapped every option a Canadian SMB has today. Each one is built for someone else.
| Option | Built for | Price | What is missing |
|---|---|---|---|
| Vanta, Drata, Secureframe | Funded startups chasing SOC 2 | $7.5K to $20K+ per year, sales call required | No human expert, no PCI or PIPEDA focus |
| vCISO platforms | MSPs, not end clients | $15K to $20K platform plus $150 to $300 per hour | Sold to the middleman |
| Manual consultants | Enterprises | $15K to $25K for a 50-person firm | Weeks of interviews, no monitoring, no benchmark |
| Alberta IT shops | Managed IT customers | Quote only | No self-serve assessment, security as an add-on |
| Jirehnexus | Canadian SMBs, 10 to 500 staff | Free to $4,750, size-based, published | Automated assessment, a named expert's signature, Canadian compliance built in |
Four tiers, each more human than the last. The free assessment is 100% automated so it can run at any volume; the expert tier is deliberately scarce. Bars show the automation share.
Twelve plain-English questions, a score on screen, the top gaps, no sign-up to begin.
Full framework assessment, 153 safeguards, MITRE-referenced recommendations, a practitioner's signature.
Gap tracker, document monitoring, periodic re-assessment and a monthly health report.
A to Z remediation with Ruphin, tracked in 15-minute chunks, balance visible in the portal.
Prices shown are the proposed size-based model, logged as pending the founder's approval. No dollar figure appears without the size tier it applies to.
Every build milestone referenced the same strategy document, so no decision was re-argued in a chat thread. Six of the twelve that shaped the design most:
Internally it is GRC automation. Publicly it is Cybersecurity Posture Analysis.
A deliberate delay window on the automated free report.
"Schools average 7 of 10. You scored 5." Anonymised, minimum-N aggregates.
Each recommendation maps to attacker techniques: recon, access, persistence, exfiltration.
First 10 minutes free for scoping; paid time auto-deducted from prepaid packages.
Six small checkers (DMARC, domain hygiene, PCI scope, insurance readiness) that each funnel into the assessment.
Automation's job is to reduce 153 safeguards to the 30 that need Ruphin. Not to replace his judgement, but to remove everything that is not judgement. The scoring core makes no external calls, LLM drafting only rewrites his notes and can be switched off, and every approval is stored with who, what version and when.
Who the buyer is, what makes them act today, the free funnel that is the whole growth engine, and why the assessment has exactly twelve questions.
Is patient data safe, and can I prove it to the college?
My processor keeps emailing me about PCI. What do I actually have to do?
The questionnaire asks for twelve controls. Which ones do we fail?
The growth engine. Green labels are automated. The one orange step is the only human touch, and it is the point.
The dip is where trust is lowest: the moment she is asked to believe a cheap assessment is a real one.
The marketing site sells, the portal delivers, the console approves. Every surface shares the same tokens, nav pattern and score language.
CIS Implementation Group 1 is 56 safeguards in 15 control clusters. Four pairs genuinely overlap and merge, which gives exactly twelve, with every IG1 safeguard covered once. The first draft ranked questions by attack coverage and silently dropped asset inventory, because nothing maps to it. Coverage zero means nothing maps to it, never that it does not matter.
Cluster coverage chooses the questions. Attack coverage ranks the findings. Two different jobs, two different tools.
Sixteen directions, one chosen. A light, calm system where green is an accent, never a background, and a mascot that looks at you. Components below are live.
The old site fell back to the OS preference and served a black "hacker" theme to every dark-mode visitor. That was its single biggest problem. The dark theme was removed entirely.
Page backgrounds are white and cool grey. Green is for buttons, ticks, meters, scores and links. Coloured gradients were removed.
On a security site, constant movement reads as anxious. Motion happens once, on scroll or on action, and respects reduced-motion. The audience skews 40s to 50s.
Never a hard-coded hex in a page. The score scale is semantic: critical to excellent, used on gauges, bands and the 18-control strip. Click to copy.
Coded from the design system. Each one is a real pattern from the site, portal or console.
Size sets the price. The framework choice is included, a second framework is +$1,500. No dollar figure without its tier.
The band changes with the score. Free results show a band, not a number, so a 3.1 never lands as a verdict.
Is there endpoint protection on every machine your people use?




The homepage hero is a four-step tour of the product inside a laptop frame, so the visitor sees the whole journey before scrolling.
Ranked by priority, then effort. Every line says what, why an insurer cares, and how big the job is.

On the live site the eyes track the pointer and the mascot stands 6px in front of the laptop frame. Both were lost twice by unrelated work, so a layout check now asserts them. Here it leans toward your cursor.

I read this before it went out. Your ISO 27001 coverage is the one to address; mostly documentation you likely already do. That's a few days of work, not a project.
Approval is a stored fact: who, which version, when. The report's value is that a named person stands behind it.
Black primary, outline secondary, green ticks for the four promises under every hero. The spring curve is the one from the design system.
"Dozens of controls in. One number out." Counters run once, when first seen.
The launch audit found tertiary text failing AA in both themes at 3.62:1. It was retuned site-wide. Every pair below is from the shipped tokens.
Scroll through the homepage. The order is a persuasion arc, not a list of features.
Four promises under the headline, the mascot, and the product tour inside a laptop. The visitor sees the whole journey before scrolling.
Three cards: twelve questions, scored against the frameworks, a practitioner approves it.
Four frameworks, 153 controls, one signature. The mascot holds the 8.2/10 report.
Ruphin's photo, four credential badges, fifteen years, and "Meet Ruphin". Added by the audit, right before the final ask.
A dark closing band that states the standard the assessment is held to, then the free assessment CTA one last time.
Hover a page to scroll it. Every page opens with a plain-English headline and the four promises.






The portal renders against five fixture states, from "nothing assessed yet" to an established client, so empty states were designed, not discovered.




Every page checked at 375, 768, 1024 and 1440 with zero overflow. The hero sizes against the real viewport height, after it was found 49px below the fold on the most common laptop.



Recorded on the build. Motion that happens once, means something, and switches off on request.
A security company's website should be simple enough to audit by reading it. Three choices that follow from that.
Twenty-four pages share tokens, nav and footer injected from one array. Open index.html and it runs. There is no dependency to patch.
Six free tools. Three are self-checks that run entirely in the browser with no fetch and no storage. Three say plainly "this check is not live yet" until the backend lands.
Three audit checks passed everywhere because they measured the wrong thing. The rule now: prove a check fails on bad input before believing its pass.
Commercial facts are not copy. Where the site said two different things, we fixed what we could and wrote the rest down.
Service page said all four, pricing said pick one. Settled: one included, a second is +$1,500.
The PDF and the meeting disagreed. $4,750 is the 25 to 100 staff tier, a size tier, never a framework price.
The quiz scores /24; every other page promises /10. Recommendation: standardise on /10 so free and paid match.
Promised in 24 places across 10 files, including the privacy policy. If no dataset exists, the copy changes and the pipeline loses a feature. Load-bearing.
From a research document to four working surfaces, with the founder's judgement protected and every unresolved commercial question written down rather than guessed.
A few other projects we've designed and built.

Visual design language and mobile app UI for aha, a Telugu OTT streaming platform, from audience research to onboarding and payment.

A 47-second promotional film for a virtual reality park: fast cuts through the park’s worlds, made for the homepage and social.

Brand identity and storefront for a fashion label: the eye mark, a four-colour system, packaging and e-commerce UI.