AI visibility audit
Outerpreneur
Outerpreneur: what an assistant can read, and what it cannot check
Partly measured on 29 September 2026. 78 live pages crawled, plus a direct request as each of 6 AI crawlers. Reviews, case studies, footer and structured data checked by hand, and every main page read in a browser, on 30 September 2026.
What this comes down to
11 things make Outerpreneur's website harder for an assistant to find, read and check, 3 of them urgent. The first is: Pages that used to work now return an error. 13 of the things a buyer would go and check fail, starting with: the legal pages are unfinished templates, the reviews cannot be verified and the case studies cannot be checked.
Read
Can an AI read this website?
This is the foundation. Before an assistant can say anything about a business, it has to be able to fetch the pages and find real text in them. None of this is about being liked or ranked: it is about whether the door opens and whether there is anything readable behind it.
- Fails, priority A+. Pages that used to work now return an error 25 page(s) returned 200 in the Wayback Machine and now return 404 or 410: /15-estrategias-de-diferenciacion-para-tu-marca-de-promocion-en-shopify, /16-differentiation-strategies-for-your-ocean-advocate-brand, /16-differentiation-strategies-for-your-shopify-advocate-brand, /20-helpful-tools-shopify-apps-for-your-ocean-startup, /33-esloganes-inolvidables-de-marcas-respetuosas-con-el-oceano, /6-avatares-completos-de-los-consumidores-conscientes and 19 more. Anything that linked to them, and any answer that cited them, is now broken.
What this means
Any link, citation or bookmark pointing at those addresses now lands on an error. That includes the pages assistants may already have read.
- Fails, priority A. The home page figures read zero to anything that does not wait The four figures on the home page are sent as "0+ PROJECTS COMPLETED, 0+ HAPPY CLIENTS, 0% CLIENT SATISFACTION, 0+ YEARS OF EXPERIENCE" and counted up by an animation that took over a minute to finish in a browser. The first crawl recorded them as zero. The targets are 100+, 50+, 99% and 7+; the CRO page says 15+ clients helped.
What this means
An assistant reads the page as sent, so it reads zero. Writing the final figures into the page, and keeping them consistent with the rest of the site, fixes both problems.
- Fails, priority A. The page about Alek is linked from nowhere /pages/shopify-theme-developer-alek-tysklind, the page that says who runs the business, has zero internal links pointing at it. Neither has the article index /blogs/shopify-theme-developer. The other orphans are 8 privacy-law pages from the cookie app (/pages/gdpr, /pages/lgpd and 6 more), which is normal. /about, /about-us, /case-studies, /case-study, /blog, /schedule, /get-started and /privacy-policy are not counted: they redirect to the home page.
What this means
Those pages are reachable only by knowing the address. Anything that finds pages by following links never reaches them. For a business sold on one person's skill, the page about that person is the one that most needs to be found.
- Fails, priority B. llms.txt is Shopify's default shopping text, not a description of the business /llms.txt and /agents.md serve the same text, which Shopify writes for every store: how a shopping bot can search products, build a cart and check out. It never names Alek, the three services, their prices or how to book a call. The theme has no agents.md.liquid or llms.txt.liquid template, so Shopify's default is what is served.
What this means
llms.txt is the one file written for AI assistants. Since May 2026 Shopify lets a store replace the default by adding templates/agents.md.liquid in the theme code editor, which covers /agents.md, /llms.txt and /llms-full.txt at once. For a one-person service business, that file should say who Alek is, what he builds, what it costs and how to hire him.
- Fails, priority B. No clear primary heading 5 pages do not state their subject in one main heading. On the home page the two H1s are the logo and the small badge "HIRE SHOPIFY EXPERT"; the real headline, "Optimization Services for Growing Shopify Stores", is not one. /pages/withdrawal and 2 articles also have several, and the case study index /blogs/shopify-projects has none.
What this means
The H1 is the page's own statement of what it is. None leaves it unstated, several leave it ambiguous.
- Fails, priority C. Pages with no meta description 7 of 78, all Shopify system pages: /policies/contact-information, /pages/withdrawal, /collections/all, /policies/terms-of-service, /policies/privacy-policy, /policies/refund-policy and /policies/subscription-policy.
What this means
The meta description is often the sentence quoted about the page. Without one, the quoting engine takes whatever text it finds first, which is frequently a menu.
- Fails, priority C. Title length off 41 of 78 pages are outside 30 to 65 characters. The case study titles are the visible case: they are cut mid-word, as in "A Bespoke Prestige Theme Build for Companiqu" and "A High-Performance Redesign for Infinite B".
What this means
Very short titles carry no information, very long ones get cut mid-phrase.
- Fails, priority C. Heading levels skip All 78 pages, for example /: h2 to h4. In the case studies the sections are emoji labels rather than headings, so after the H1 there is no H2 at all.
What this means
Headings are how a machine works out which parts of a page belong to which. Skipping levels leaves that hierarchy less clear than it could be.
- Fails, priority C. Images with no alt text 49 image(s) across 34 page(s).
What this means
Alt text is what an image says to a screen reader and to anything reading the page as text. Missing, the image says nothing.
- Fails, priority C. Supporting schema missing Not found: Service, FAQPage. Person is covered under Trust.
What this means
Service, FAQ and Person blocks are how the specific things you do, the questions you answer and the people who do the work become readable rather than inferred.
- Fails, priority C. Every case study carries two structured-data blocks that disagree Each case study has two Article blocks with different ids and different dates. On the Made to Order case study one says published 14 September, the other 11 September. Neither names the client. And in the page trail, the apostrophe in "Coluri's" is stored as the code ', so a machine reads the client's name as Coluri's.
What this means
When two descriptions of one page disagree, a machine has to pick one. Naming the client as an Organization with its website is what ties the work to a business anyone can look up.
- Not checked. Whether the text is in the HTML before JavaScript runs not checked Comparing what a fetch returns against what a browser ends up with needs the page opened in a real browser, and that was not done on this run.
What this means
Some sites send an almost empty page and then draw the content with code once it is open in a browser. People see a full page. An assistant that only reads what was sent sees the empty one.
18 checks pass
- Passes. Those crawlers actually get in Asked for the homepage as each of the 6 in turn, and every one was let in with a 200. This is a separate check from robots.txt: a rule at the CDN can block a crawler the file allows.
- Passes. robots.txt is there and it reads correctly Fetched and parsed, not just found: 6411 bytes of real directives, 4 user-agent group(s), 81 disallow rule(s), and it points at the sitemap.
- Passes. The sitemap covers the pages that matter. 73 addresses declared. The live pages it leaves out are Shopify system pages (/collections/all, /search and the policies).
- Passes. Titles are not repeated. The 10 pages first reported with the same title were the home page and 8 addresses that redirect to it.
- Passes. Descriptions are not repeated. Only /search reuses the home page's description, and it is a Shopify system page.
- Passes. A sitemap is published, so the pages are declared rather than only linked.
- Passes. Every live page has a title.
- Passes. Every page declares which address is the real one.
- Passes. Shared links preview properly.
- Passes. No page is too thin to answer anything.
- Passes. The site carries structured data, so its facts are machine-readable.
- Passes. The business itself is described in structured data.
- Passes. Structured data reaches every live page. 72 of 72 HTML pages in the sitemap carry it. The 73rd address, /agents.md, is a text file Shopify generates.
- Passes. Every structured-data block parses.
- Passes. A made-up address returns a real error instead of a page.
- Passes. The error page offers a way back into the site.
- Passes. No tracking tag is double-counted.
- Passes. Pages link to each other, so the site can be walked.
Trust
Can a buyer check what this site says?
A site can be perfectly readable and still not be believed. This step compares what the site claims against what a stranger can verify: a number behind every claim, a person behind the business, and the same details wherever the business appears.
- Fails, priority A+. The legal pages are unfinished templates /policies/terms-of-service still contains 7 template gaps: [INSERT TRADING NAME], [INSERT BUSINESS ADDRESS], [INSERT BUSINESS PHONE NUMBER], [INSERT BUSINESS REGISTRATION NUMBER], [INSERT VAT NUMBER], [LINK TO REFUND POLICY] and [LINK TO PRIVACY POLICY]. /policies/refund-policy is Shopify's text for physical goods ("unworn or unused, with tags, and in its original packaging") and still reads [INSERT RETURN ADDRESS]. The trading name, address and VAT number are already on /policies/contact-information.
What this means
These are the pages a buyer reads before paying 2,500 to 6,000 EUR. Filling the gaps takes minutes, and the refund policy should describe what happens when a service, not a parcel, goes wrong.
- Fails, priority A+. The reviews cannot be verified The Judge.me app holds 13 reviews: 12 on the store build service, 1 on CRO, none on speed. All 13 are marked not verified and collected from another provider, which means imported. The theme hides that label with CSS, so a visitor cannot see it. The site claims 12 reviews in its structured data, 21 on the service pages (including the speed page, which has none) and 25 on the About page, and the home page badge says 4.9. Jens Vogel appears 3 times, twice on 31 March 2023. One home page testimonial is still signed with the template placeholder "Customer name, Verified buyer".
What this means
A rating is only worth what a stranger can check. Imported reviews are fine to show; hiding that they are imported, and giving three totals, is what makes every other claim look less certain. The fix: show the source label Judge.me already provides, state one total everywhere, remove the duplicate and the placeholder, and point to the verified Upwork reviews (see below).
- Fails, priority A+. The case studies cannot be checked 41 case studies, and none links to a client website. 37 follow the same template; the newest 9 use emoji headings and a Key Stats block, and 4 of those say "Exact performance metrics pending". The four I read in full are in the section below.
What this means
For a Shopify developer, the case studies are the proof. They are where an assistant asked who to hire, and a buyer comparing developers, look first.
- Fails, priority A. The rating stars do not show On the home page every testimonial shows five empty outlined stars, which reads as zero out of five. On the three service pages the Judge.me stars, next to "5.0 out of 5 stars based on 21 reviews" and under "Satisfaction guarantee", show as white boxes. Only /pages/reviews shows the stars correctly.
What this means
A broken star is worse than no star: it is the first thing a visitor checks, and it looks like a bad score or a broken site.
- Fails, priority A. The Upwork profile that vouches for Alek is not linked Alek's Upwork profile (upwork.com/freelancers/~0174b0f6a408628337) shows Top Rated Plus, 100% Job Success, 5.0 from 10 reviews, 34 jobs and 2,300 hours, based in Geneva, with the same portrait as the site. The site never links to it. What it links instead are profiles the business runs itself: his personal LinkedIn and Outerpreneur's Facebook, Instagram and X.
What this means
Upwork reviews can only come from clients who paid a contract, which makes them the strongest proof this business has. Link the profile from the footer, the reviews page and the structured data, and show its rating in place of the 4.9 and 5.0 figures the site states now.
- Fails, priority A. The Shopify Partner claim cannot be checked The home page headline block says "Shopify Agency Partner for today's Leading DTC + B2B Brands", and a Shopify Partners logo sits at the top of the home page and of all three service pages and in the footer of every page. None of them links anywhere. The footer file is named ShopifyPlus_Secondary_Inverted.png. I found no public profile in the Shopify Partner Directory under Outerpreneur or Aleksander Tysklind.
What this means
It is the site's lead credential, on every page, and a buyer has no way to confirm it. Any member of Shopify's Partner Program may show the logo, so the claim may well be true; what makes it checkable is a Partner Directory profile, linked from the badge. If there is none, the badge and the word Agency Partner promise more than can be shown.
- Fails, priority B. The site speaks as one person and as a team 24 of 60 pages mix I and we. About 500 uses of I and my against 160 of we, our and us. The store build page does it inside one sentence: "With our expertise in Shopify development, I can create". The same testimonial reads "the Outerpreneur team, they" on the home page and "Alek, he" on the reviews page, and is signed CEO at Carvonis Group in one place and CEO of Dlish in the other.
What this means
An assistant describing the business has to decide whether Outerpreneur is a freelancer or an agency. The structured data says an organisation, the About page says one person, the footer says we. Picking one voice, most likely I, settles it.
- Fails, priority B. The About page puts Alek in Asia; everything else says Geneva The About page lists "2024 - Present, Remote / Asia" and "Remote (Asia)" under Places I've Called Home. The structured data, the contact information, the opening hours and the Upwork profile all say Geneva. The street number differs too: Cours de Bastions 13 in the structured data, no number on /policies/contact-information.
What this means
Where a business is based is one of the first facts an assistant repeats. Two answers on one site means it may repeat either.
- Fails, priority B. The GitHub links on the About page return an error Three project cards on /pages/shopify-theme-developer-alek-tysklind carry a GitHub button. Superbra points to github.com/outerpreneur/Superbra-3.0.git and WeatherTech to github.com/outerpreneur/Maranello-5.0.git; both return 404, and neither is among the 89 public repositories on that account. The Knockout H2O button is labelled GitHub but opens the case study. The public repositories are mostly Frontend Mentor exercises, with no client Shopify theme among them.
What this means
The About page is where a buyer goes to find out who does the work, and a proof link that lands on an error is worse than no link. Either make the two repositories public, or remove the buttons and label the Knockout H2O one View case study.
- Fails, priority C. Alek is named only as the author of articles The 50 articles name Aleksander Tysklind as author in their structured data, and the footer links to linkedin.com/in/aleksander-tysklind. The home and About pages carry no Person block, and the author block does not point to LinkedIn or to his public GitHub (github.com/outerpreneur).
What this means
Structured data is the machine-readable version of facts already on the page. Linking the person to his profiles is what lets an assistant tell the founder from a name in a testimonial.
- Fails, priority C. The YouTube icon opens Shopify's channel The footer's YouTube icon links to youtube.com/shopify, the theme's default, not to an Outerpreneur channel.
What this means
A profile link that goes to someone else's channel is a small thing, and it is the kind of small thing a careful buyer notices.
- Fails, priority C. The contact form does not ask which service The contact form asks for name, email, phone and a message. The other forms on the site are the newsletter sign-up in the footer and the search box. The Book a call page uses a cal.com calendar, which does not ask either.
What this means
With three services at three prices, every enquiry starts with a question the visitor already answered by choosing a page. A drop-down with the three services fixes it.
- Fails, priority C. Typos and mismatched names in the copy Iwill, Multiople, 5-strep process, From our Porfolio, Have question?, Outerpreneur Tea, Outerprener, expectatitons, whom pays, He is cares. The same reviewers appear under different names: Mario Leao and Mario Leang, Emilie Disney and Emilie Disner, Floor V.P and Floor Van de Pavert. The speed page lists FID, a metric Google retired in 2024, and shows a 57-point checklist with the 115-point figures of the CRO page.
What this means
Each one is small. Together, on a site selling attention to detail, they are what a careful buyer notices first.
- Not checked. Whether a licence number is published not checked This business does not need a licence to operate, so there is no number to look for.
- Not checked. Whether the profiles match the site, field by field not checked Comparing name, address, phone, hours and services across https://www.linkedin.com/in/aleksander-tysklind/, field by field, needs each profile opened in the browser, and it was not done on this run.
What this means
Where a name, address, phone or list of services differs between a website and its profiles, the difference is what gets noticed, and it makes every other claim look less certain.
1 check passes
- Passes. A price is published where a visitor can read it Found on 24 page(s), for example /products/shopify-store-build: 2,500, 4,500 and 6,000 EUR.
Recommend
Would an assistant recommend this business?
Read and Trust are facts about one website on one day, and anyone with a browser can check every line above. Whether an assistant actually recommends a business is a different kind of question: one answer proves nothing, because it can change tomorrow. Answering it properly means asking the questions your buyers ask, week after week, watching who gets named, and telling a real shift from noise. That is the paid work, it is not part of this report, and this report is complete without it.
The case studies
They live under /blogs/shopify-projects. /case-studies, the address a buyer would type, redirects to the home page.
I read 4 of them in full:
| Case study | Client store today | What is missing |
|---|---|---|
| Made to Order, Swiss womenswear | Not named | Client name or link, any screenshot of the fabric picker, any number |
| Never Truly Sold Out, Coluri | coluri.com is live; the case study’s address says coluris, a colouring app | Screenshot of the new message, the pre-order count before and after |
| From Figma to Storefront, Companique | companique.com is closed behind a password | Everything a closed store cannot show; the designer, John Bresciani, is linked: the one outside link in all 41 |
| App-Free Customization Hub, Infinite Battery | The store (Gouach) is closed behind a password; the case study never says Gouach | Any speed figure behind “drastically improved site speed” |
What each case study needs, in order:
- The client’s name and a link to the live store. If the store is not live, publish the case study when it is.
- Three to six screenshots, with a before and an after.
- One real number with its source: a PageSpeed score, apps removed, pre-orders, conversion rate, days to launch.
- A quote from the client, with name and role.
- No Key Stats block when there are no stats.
Fewer and stronger beats many and alike. Ten case studies that meet that list, on a real /case-studies page and linked from each service page, would do more than the 41 there are now.
How this was measured
Eight addresses that redirect to the home page were counted as pages on the first pass; the figures above leave them out. On 30 September 2026 I checked the reviews through Judge.me’s public widget, read four case studies and the client stores they name, read the structured data on every page in the sitemap, read every main page top to bottom in a browser, and read Alek’s Upwork profile. Every check above is a fact from that crawl or from a request made to the site, and each carries the evidence it came from.
Nothing needed an account, and nothing was measured that a visitor cannot see.
Nothing here was measured that is not publicly visible, and no account access was used. The site read for this report was https://outerpreneur.com, observed on September 29, 2026.
Outerpreneur: if you would rather this came down, email support@areyouvisible.co and it is removed the same day, no questions asked.
