The client is a financial advisory firm specialising in credit access, regulatory compliance, and growth advisory for mid- and small-scale manufacturing enterprises.
The advisory category is information-driven. Business owners come looking for current, jurisdiction-specific guidance on schemes, eligibility, compliance obligations, and sector opportunities.
The firm's commercial relationship with a client typically begins with that owner finding the right answer to a specific question on the website. The platform had to carry that weight. Editorial credibility and discoverability were the primary commercial assets.
The inherited platform was struggling on discoverability, editorial velocity, and forward-readiness for AI-mediated search. The structural issue underneath was the same: content was treated as a developer concern, not as the product.
Long-tail organic search is the dominant customer acquisition channel in this category. The existing platform was slow and missed the technical SEO foundations needed to rank on the specific scheme- and sector-level queries that mattered.
Publishing required developer involvement. Regulatory updates that should have been live within hours sat in queues for days.
AI-mediated search was beginning to redirect a meaningful share of advisory queries away from traditional Google results. The platform had no surface area for AI assistants to read or cite.
Underneath those three was a structural problem: the architecture treated content as a developer concern. The firm needed an architecture that treated content as the product.
We rebuilt the platform around a simple operating principle: content is the product. The architecture had to make content fast to publish, structured enough to query, performant enough to rank, and readable by both search engines and emerging AI assistants.
Each technical choice was tied to the firm's commercial model: organic discovery, trusted answers, fast updates, and low operational risk.
Angular with Server-Side Rendering
Contentful Headless CMS
Static Deployment via Netlify
AI-Augmented Editorial Workflow
AI discoverability: a low-cost early bet
Delivers fully-rendered HTML to crawlers and first-time visitors.
Faster first-paint and better performance support rankings and user experience.
Every millisecond of performance improves the ability to rank for high-intent queries.
Performance that drives discoverability and growth.
Models eligibility, sectors, rates, timelines, versions, and queryable fields.
Editors publish independently while engineers ship features with no blockers.
Structured content enables filtering, reusable components, and programmatic SEO.
Structured content is the foundation of the platform.
No runtime database, application layer, or admin login surface.
Smaller attack surface and simpler compliance for fiduciary responsibility.
Contentful publish events trigger instant build and deploy.
Secure, fast, and current.
Every draft goes through a 3-step AI pipeline:
AI generates content from verified source documents.
Refines tone, clarity, and reader experience.
Cross-checks factual claims against the source for accuracy.
Every output is reviewed and approved before publishing.
The pipeline expands what editors can cover; humans stay accountable for what gets published.
Structured file that allows AI assistants to read, understand, and cite content.
A low-cost early investment in an emerging surface with high future potential.
Small implementation effort for long-term strategic advantage.
Future-ready discoverability for the AI search era.
A content-led platform built around technical performance, structured content, editorial agility, and AI readiness.
Two things needed correction in the first weeks after launch. Both were cheaper to fix at week 4 than at month 6. Launch was the start of the build, not the end.
Our initial content models were too generic: body fields where structured fields belonged. This limited the filtering and landing-page generation we had designed for. We refactored the schema into explicit eligibility, geography, timeline, and documentation fields.
Editors initially over-relied on AI first drafts, treating generated content as closer to final than it should have been. We tightened the source-grounding workflow and made fact reconciliation a mandatory gate, not a recommended step.
Within weeks of launch, the firm moved from developer-gated publishing cycles to same-day editorial releases. Publishing throughput improved roughly 4x over the previous workflow. Technical performance improved: Lighthouse scores moved from 54 to 91, with downstream improvement in long-tail organic visibility on the queries the firm targets. Organic traffic grew 56% in the 3-month window after launch.
Beyond any single metric, the firm now owns a publishing engine rather than a brittle marketing website. The editorial team operates without engineering as a blocker. Regulatory updates land in hours, not days. The platform architecture aligns with how this category actually generates commercial relationships: through editorial credibility, surfaced via search and AI assistants, and converted through trust.
Regulatory updates that previously waited in developer queues could go live within hours.
Publishing throughput improved roughly 4x over the previous workflow.
Technical performance improved, strengthening Core Web Vitals and long-tail organic visibility.
Organic traffic grew 56% in the three-month window after launch.
The editorial team operates without engineering as a blocker.
The platform exposes structured surfaces to traditional search and emerging AI assistant retrieval.