We integrated AI content generation into Check This Car’s Laravel CMS, connecting motoring news feeds to structured article drafts and an editor-led publishing process. The development brings source discovery, drafting, revision and publication into the website’s existing administration tools.
Keep a motoring website useful between vehicle checks.
Check This Car helps UK motorists explore vehicle information. Alongside its checking tools, useful articles can help readers understand the issues behind a purchase: vehicle history, recalls, MOTs and the questions to ask before choosing a used car.
Keeping that content current takes time. Someone needs to spot relevant stories, understand why they matter to the audience, prepare a clear article and get it into the website’s publishing system. We brought AI into that process to help with the preparation while keeping editorial decisions with the people running the site.
Find the stories that fit the audience.
The workflow gathers items from configured news feeds and checks for motoring relevance. Topics such as used cars, vehicle history, MOTs, recalls and vehicle fraud help identify candidates for drafting. Source references are retained so the editor can trace a draft back to the story that prompted it.
Generation can run through the scheduled process or be started manually by an administrator. Both routes use the same workflow, with a shared lock to prevent overlapping runs. The system records which source items have produced drafts, giving administrators a clearer view of what has been processed.
Give AI a specific editorial job.
The drafting instructions ask for an original article based on the facts supplied in the feed item, written in British English for UK used-vehicle buyers. They ask the AI to explain why the story matters, offer useful context and suggest sensible checks, with attribution and a link to the source.
The result is structured for the existing CMS: an article title, a meta description, a category and formatted HTML with useful headings and a practical checklist. Generated HTML is sanitised before it is saved. The content enters the system as a Blog Draft, ready for review.
Help editors revise, not just generate.
Editors can also work with AI from within the content editor. They can provide instructions, supply an existing draft for revision and use publishing rules that request a consistent HTML structure. This lets an editor ask for a clearer explanation, a more useful angle or a better-organised article without moving the work between disconnected tools.
The publishing rules ask for relevant facts and source links to be preserved and prohibit invented claims, dates and statistics. These are instructions to the model, so editorial checking still matters. Editors need to verify the article’s facts, source attribution and links, and make sure the result helps the intended reader.
Fit the process around the website.
The content workflow works with Check This Car’s existing administration and publishing tools. Draft generation, article editing, categories and metadata sit within the same system, rather than creating another content store for the team to manage.
Source-item tracking and per-item generation results help administrators see what has been drafted and which items need attention. A problem with one generation request can be recorded without losing the rest of the work. The manual route gives the team control over when to prepare another set of drafts.
Practical AI, with an editorial purpose.
Laravel development behind the AI content workflow.
The integration uses a Laravel service to discover RSS and Atom feed items, score their motoring relevance and prepare a bounded set of drafts. Each source item retains its reference and processing status, allowing the application to associate the generated draft with the story that prompted it.
The model is asked for a structured response containing a title, meta description, category and HTML. The application checks the response structure, generates a unique URL slug and sanitises the HTML before saving it as a Blog Draft. Scheduled and manual runs share a lock to prevent them overlapping, and generation failures are recorded against the relevant source item.
Within the editor, an administrator can supply an existing draft and new instructions for revision. The application validates the request and returns the result to the editing workflow. These are application-level controls around the model, alongside the editor’s checks of facts, sources and links.
This is AI integration for existing business software applied to publishing. Our wider API and systems integration work brings the same attention to how inputs are gathered, responses are handled and failures are surfaced.
Content automation with an editorial purpose.
This project brings AI into a repeatable content process: finding relevant material, preparing a useful first draft and supporting revision. The aim is to give editors a better starting point and keep the work connected to the tools they already use.
Content quality comes from relevance, accuracy and a clear point of view. AI assistance gives the team another way to develop that content, while editors remain responsible for what reaches readers. Time spent assembling a first draft can become time spent checking the detail and making the article more useful.
Turn your knowledge into content people can use.
Explore our hospitality AI integration case study and AI development for legal-register workflows for other examples of connecting business information to a focused, reviewable output.
Whether your starting point is a news feed, specialist knowledge or an existing library of material, we can help you build an AI-assisted workflow around your audience, your editorial standards and your CMS.
Talk to Impact about AI for content creation and publishing.
The hero image is an AI-generated illustration of an editorial workspace, not a photograph of Check This Car staff. The workflow diagrams illustrate the process and contain no customer data or performance results.