Insights

From the field, not from theory.

AI AgentsReliabilityProductionMonitoring

Every health check was green while the agent loop had been dead for 9 hours

Our monitoring said all systems healthy while an AI agent silently stalled overnight. Here is how we rebuilt health checks for LLM workloads that fail in ways HTTP 200 cannot see.

Custom SoftwareSecurityProduction

We shipped a strict CSP to a live SME platform and broke checkout for 6 hours

Hardening a server-rendered platform with Content-Security-Policy looked like a one-line header change. It took down payment flows in three browsers and taught us a rollout pattern we now reuse.

Generative AIEvalsRAGProduction

Our LLM judge agreed with humans 61%: calibrating evals for an assistant

We shipped an LLM-as-judge eval pipeline for a grounded assistant and trusted it for weeks. Then we hand‑labelled 200 cases and found the judge was wrong four times out of ten. Here is how we fixed it.

Semantic SearchRAGpgvectorProduction

Embeddings alone failed on part numbers: what hybrid search fixed for us

Pure vector search scored 61% on queries containing SKUs and error codes. Adding BM25 alongside pgvector lifted retrieval accuracy to 89% without touching the embedding model.

Process AutomationReliabilityProductionArchitecture

Queue fan-out vs fan-in: the deadline we almost missed

A 14-job pipeline looked fine in staging. In production, fan-in ordering cost us a 6am client deadline and taught us three rules we now enforce.

GEOSEOMeasurementSchema

We tracked 12 schema types across 90 days: only 3 moved AI Overview citations

After three months logging every FAQ, HowTo and Product schema on client sites, the citation lift came from three types nobody talks about.

AI AgentsProductionReliability

Agent retry storms: when exponential backoff meets shared state

Three agents retried the same failed job in lockstep and cost us €40 in 12 minutes. Here is the jittered, state-aware retry pattern we shipped instead.

AI AgentsProductionPerformance

We cut LLM timeouts from 30s to 4s and lost no quality

A p95 latency spike on a customer-facing agent forced a brutal timeout decision. Here is the data, the trade-off, and what we measured before and after.

Custom SoftwareArchitecturePerformanceSME

We stopped shipping SPAs for SME platforms: what server-rendering actually changed

Across four SME platforms we rebuilt as server-rendered apps, time-to-interactive dropped by roughly 60%, hosting bills halved, and one whole class of bugs disappeared. Here are the numbers and the trade-offs.

Generative AICost ControlProductionArchitecture

Prompt caching is not a free lunch: the budget rule we had to add

We turned on prompt caching for a grounded assistant and the bill tripled before we noticed. Here is the budget rule we now enforce on every LLM copilot.

Data IntelligenceRAGpgvectorSemantic Search

When pgvector recall falls apart: what we saw at 500k rows

Our RAG system quietly degraded as the corpus crossed half a million embeddings. Here is the exact threshold, the metric that caught it, and the fix that did not cost us a rewrite.

AutomationReliabilityProductionArchitecture

Idempotency keys: the boring pattern that stopped our pipelines double-billing clients

Retries are easy to add and brutal in production. How one deduplication key per job stopped duplicate invoices, double-sent emails and a very awkward client call.

GEOSEOMeasurement

We measured llms.txt for 90 days: here is what GPT and Perplexity actually fetched

We instrumented llms.txt on nine client sites for 90 days. The honest numbers: crawl rate, citation lift, and the bots that never showed up.

AI AgentsCost ControlProduction

How we stopped a runaway agent from burning €180 in a weekend

An autonomous content agent got stuck in a retry loop and burned through its monthly budget in 48 hours. Here is the guardrail stack we built to make sure it never happens again.

AI AgentsSEOAutomationCase Study

How a biomass boiler manufacturer went from one static page to a 6-language AI platform

Tatano sells in Luxembourg, Belgium, France and Switzerland. In six months: four country domains, six languages, a 24/7 AI chatbot, a daily autoblog and generative video campaigns — all operated autonomously.

AutomationSocial MediaAI AgentsCase Study

4 publications a day, zero manual effort: how we built an autonomous social pipeline

TikTok, Instagram, Facebook, YouTube — four posts a day, every day, across all platforms. No editor, no scheduler, no agency. Here is the architecture that makes it run.

AutomationAI AgentsCase Study

Le Podcast IA: controversial daily topics discovered by AI, produced and published end-to-end

Every night at 03:00 UTC, discover_daily_topic.py finds the hottest controversial AI story (<48h) via Google Trends, YouTube audience signals and Grok web search — then produce_daily.py turns it into a full episode on @lepodcastia without human intervention.

SEOGEOAutomationContent

From zero to daily organic content: how an AI autoblog compounds over time

An article a day, published automatically, indexed within minutes, distributed across every channel. The architecture behind a content engine that runs without a content team.

AutomationSocial MediaAI AgentsCase Study

Publishing a trance music catalog to TikTok, Facebook and Instagram — automatically, every day

795 MB of licensed trance tracks. Two YouTube channels. TikTok, Facebook, Instagram. A hook bandit algorithm. All driven by TikTok Studio and the Blotato API — zero manual publishing.

AI AgentsB2BChatbotCase Study

A B2B lead chatbot in 7 languages, live in 4 countries — built in a weekend

Optimus answers product questions 24/7 across tatano.lu, tatano.be, tatano.fr and tatano.ch in French, English, Dutch, German, Italian and Luxembourgish. 1.3-second response time. Here is how it was built.

Custom SoftwareAI AgentsCase Study

abbysconsult.space: a consulting firm's website that never sleeps

A premium showcase for an independent Belgian strategy consulting firm — editorial design, sub-200ms TTFB, zero technical debt at launch, and an AI assistant powered by NVIDIA NIM running on the same server.

SEOGEOAutomationCase Study

matthieupesesse.com: a personal media platform that publishes, distributes and optimizes itself

One Next.js platform. One article per day, auto-generated. Newsletter, social and IndexNow on every publication. GEO audit from 66 to 90+. Zero hours per week of content operations.