Cadman Training: DeepSeek
AI Course Advisor

An editor lands on a training site knowing exactly what they want to get better at, and then stalls. The catalogue spans PerfectIt, EndNote, Microsoft Word, and referencing; the home page is busy with tabs that no longer earn their place; and there is no quick way to ask “which course is right for me?” without sending an email and waiting for a reply. A visitor who came ready to learn leaves unsure.

Cadman Training is an online training business that teaches editors to master the professional tools of their trade. This project revamps its website to be warmer and easier to move through, and adds an AI Course Advisor powered by the DeepSeek API. The advisor turns the catalogue into a personal recommendation: a visitor answers a few short questions, and it points them to the right course or courses with direct links, grounded in Cadman Training’s own offering.

Put simply, the site was reordered around the visitor and given a guide that never sleeps. The home page leads with the trainer’s own story, testimonials from working editors, and clear reasons to learn here; every subject page moves to one welcoming layout; the rich Word offering is untangled into clear pages; and the DeepSeek advisor sits across all of it, matching each unsure visitor to a course in a few questions instead of a wait.

The Starting Point: Turn a Catalogue Into a Guided Path

An expert trainer’s site should turn curiosity into enrolments, but a growing catalogue and a cluttered home page were getting in the way. That defined the work, and most of it came down to clarity and to giving each visitor a confident next step.

• Help unsure visitors choose. Editors arrive wanting a specific tool or skill but cannot always tell which course fits, and unsure visitors leave, so the site needed a way to recommend the right one on the spot.
• Make subject pages welcoming. Pages explaining each piece of software read like documentation rather than an invitation, which undersold friendly, practical courses.
• Cut the clutter on the home page. An unused blog tab, a webinar line, and a generic trainer tab distracted from what actually converts: the trainer’s story, real reviews, and the reasons to learn here.
• Strengthen and broaden the social proof. Testimonials that did not reflect a global audience missed a chance to build trust with editors everywhere.
• Untangle the Word offering. A main course, mini-courses, and coaching squeezed into one section were hard to understand and hard to find in search.
• Keep every answer grounded in real courses. A guide that invented courses would mislead visitors, so any recommendation had to come from the actual catalogue and link straight to it.

The Build: A Warmer Site With a Guide Built In

The revamp reshaped the pages around the visitor and layered one AI feature through the whole journey. Grounding that advisor in the course catalogue so it recommends courses that exist, with direct links, rather than improvising, is the substance of the custom DeepSeek app development behind the site. The work ran in a clear order:

1. The home page was decluttered and reordered: an About Me section with a photo replaced the generic trainer tab, “why learn with Cadman Training” moved higher, and the blog tab and webinar line were removed.
2. The testimonials section was redesigned with reviews from editors around the world, and the tech-tools section shifted its focus to the trainer’s YouTube channel while keeping a blog link.
3. Social icons for Facebook, LinkedIn, and Instagram were added, and the editing-society line became a simple, general offer: a 25% discount for society members.
4. Every subject page, PerfectIt, EndNote, and the rest, moved to one warm, consistent layout: an image on the left, clear text explaining the software, a summary of the relevant courses with clickable links, a testimonial, and thumbnails of related courses at the bottom.
5. The Word offering gained a short intro page explaining the main course, mini-courses, and coaching, then split into three dedicated pages using that same structure so each one is clear and individually findable.
6. The AI Course Advisor was built across the site: a visitor answers a few short questions about the software they use and what they want to achieve, and it recommends the right course or courses with direct links.

The advisor only ever recommends courses that exist, pulling names, descriptions, and links from the catalogue so it never sends a visitor toward something Cadman Training does not offer, and it answers follow-up questions about each course it suggests. DeepSeek also helped draft and structure the refreshed copy for the subject and course sections, which the trainer reviews and approves before anything goes live. That approval step matters: the site stays in the trainer’s voice, and the AI does the heavy lifting rather than the deciding.

Technical Architecture

Cadman Training is a content-driven course website with one AI feature layered in. The hard part was not the pages themselves; it was making the site feel warm and consistent while adding an advisor that recommends courses accurately and links straight to them. Each layer supports clarity, trust, and a guided path to enrolment.

• Experience layer. A refreshed home page, one welcoming subject-page template, the Word intro plus three dedicated pages, and clearer course pages give the whole site a single, inviting structure.
• DeepSeek AI Course Advisor. The DeepSeek API asks a visitor a few short questions about software, experience, and goals, then recommends the right courses and answers questions about them. That DeepSeek API integration is the one AI feature the whole journey leans on, and wiring a language model cleanly into an existing site sits at the core of our AI integration services.
• Advisor grounding. The advisor reads from the course catalogue and knowledge base, so its recommendations and links come from courses that exist rather than guesswork.
• WordPress CMS. The site runs on WordPress with a page builder, so the reusable subject and course templates give one consistent layout and the trainer can update most pages herself.
• Content and media. A structured catalogue, redesigned global testimonials, and YouTube and social embeds carry the courses, the social proof, and the trainer’s presence.
• Analytics and performance. GA4 tracks journeys, advisor use, and enrolment clicks, and caching with on-page SEO keeps the subject and course pages fast and findable.

The join worth engineering carefully was between the advisor and the catalogue, because a recommender that drifts off the catalogue does more harm than no recommender at all. Keeping every answer grounded in courses that exist, with a direct link to each, is what makes the DeepSeek AI integration something a visitor can trust rather than second-guess.

Challenges Solved

The harder problems were the ones that decide whether an interested visitor enrols or drifts away. Helping unsure visitors choose was solved by an advisor that asks a few questions and links to real courses, so a visitor gets a confident answer instead of guessing. Keeping recommendations accurate was solved by grounding the advisor in the actual catalogue, with the trainer’s own content behind every reply. Dense subject pages were solved by one reusable, welcoming layout that turns documentation into an introduction. The tangled Word offering was solved with a short intro and three dedicated pages. And a site that felt generic was made personal again by leading with the trainer’s story, a photo, redesigned global testimonials, and refreshed social links.

What ties it together is a guided path rather than a longer catalogue, which is what online course and training businesses need when the difference between a browse and an enrolment is one confident recommendation at the right moment.

The Result

The outcome is a site that feels like learning from a real person and guides each visitor toward the right course. The home page leads with the trainer and the reasons to learn here; every subject page shares one welcoming layout; the Word offering is clear and findable across four pages; and the DeepSeek advisor turns “which one is for me?” into an instant, grounded recommendation with a direct link. Because the templates are reusable and the CMS is manageable, the trainer can keep most of it current herself.

The build works because the advisor is grounded in the catalogue and the whole site was reordered toward what converts, rather than bolting a chatbot onto a cluttered page. If you are planning a guide like this, you can hire agentic AI developers who have grounded a recommender in a working catalogue, or tell us about your courses and your visitors and we will map them to a DeepSeek advisor that recommends the right one.

Frequently Asked Questions

What did the Cadman Training project involve?

It was a website revamp for an online training business that teaches editors to use tools such as PerfectIt, EndNote, and Microsoft Word. The site was made warmer and easier to navigate, the subject pages moved to one consistent layout, the Word offering was split into clear pages, and a DeepSeek-powered AI Course Advisor was added to guide each visitor to the right course.

What does the AI Course Advisor do?

The advisor helps a visitor find the right course in a few short questions. It asks about the software they use, their experience, and their goals, then recommends the right course or courses with direct links and answers follow-up questions about them, so an unsure visitor gets a confident next step instead of leaving.

How does the advisor avoid recommending the wrong course?

It is grounded in Cadman Training’s own catalogue. The advisor pulls course names, descriptions, and links from that catalogue, so it only ever recommends and links to courses that exist, with the trainer’s own content behind every answer.

Which AI powers the site, and who approves the content?

The AI Course Advisor and the content drafting both run on the DeepSeek API. DeepSeek helps draft and structure the refreshed copy for the subject and course sections, and the trainer reviews and approves everything before it goes live, so the site stays in her voice.

Can the trainer maintain the site without a developer?

Yes. The site runs on WordPress with a page builder and a reusable subject-page template, so new subject pages are quick to add and the trainer can update most of the site herself. GA4 tracks journeys, advisor use, and enrolment clicks so she can see what is working.