Browse all articles from roiscale
The maintenance AI was trained on twelve years of ticket logs, but it missed the thing every veteran technician knows — the sound the compressor makes the week before it fails — and what's lost when embodied expertise is excluded from industrial AI is harder to recover than the data scientists assume.
Site search is the most under-optimized surface in e-commerce. I swapped a traditional search for a Perplexity-backed agent on a Shopify store. Conversion rose 22%. Here is the setup.
The OWASP LLM Top 10 reads like it was written for enterprise. I ran all ten against a side-project agent I had running in production. Three hit. One was embarrassing. Here is what I fixed and the code.
She spent two years training a credit-adjudication AI — then the bank cut her role — and the story of who benefits when human judgment is codified is also the story of who is expected to do the training without being paid for it.
Ambient clinical documentation is being positioned by vendors as a physician-retention and burnout-reduction tool — a compelling narrative that is, nonetheless, the wrong frame for finance leadership.
Gig delivery drivers used to choose their routes, but now the algorithm chooses for them — monitoring every stop, every second, every customer complaint — and what happens to workers when the algorithm becomes the supervisor is a story about control that the independent-contractor classification was never designed to address.
AI-assisted content production has structurally altered the marginal cost of premium media, reducing video post-production cost per minute from $1.80 to $0.24 and localization cost by 58%. The strategic implication is not cost reduction — it is capacity reallocation.
Synthetic data generation is the 'just add water' of fine-tuning advice. I tried it at scale for a real classification task. The 30% synthetic sweet spot is real. Here is what goes wrong when you push past it.
My city's 311 system routes everything to a human dispatcher. I scraped 18 months of requests, trained a classifier, and built a chat layer. The routing accuracy is better than the baseline. The city wants in.
Construction firms that deployed AI for project scheduling, estimating, and document management are reporting working-capital benefits that substantially exceed the direct productivity gains that motivated the original investment.
A financial-services client was running a multi-stage RAG pipeline with fixed-size chunking, single-pass vector retrieval, and no reranking; per-query cost was $0.47 and accuracy on domain evaluation was 61%. Rebuilding the pipeline with semantic chunking, hybrid BM25-plus-dense retrieval, and a cross-encoder reranker dropped per-query cost to $0.11 and raised accuracy to 79%.
RFPs are where consultants go to die. I gave our team an AI that reads the RFP, finds the matching case studies, and drafts the response. Win rate up 19%, proposal-writing hours down 70%.
