ArticlesAug 20267 min read
Why AI pilots stall before production
Most enterprise AI pilots work. Far fewer reach production, and the reason is almost never the model.
Data, Analytics & AI practice
Insights
What we think, and why we think it.
Our current position on enterprise technology, and the reasoning behind it.
Latest
5 published
ArticlesJul 20266 min read
Reducing testing time without reducing coverage
When a regression pass takes two days, the instinct is to automate more of it. The faster answer is usually to move it to a different layer.
Quality Engineering practice
Executive guidesJun 20269 min read
Replatform, refactor or rebuild: deciding on evidence
Three legitimate answers to the same legacy problem, with very different costs. An executive guide to choosing between them on evidence rather than preference.
Product & Platform Engineering practice
Industry points of viewAug 20266 min read
Modular content and the medical review constraint
In life sciences commercial operations, review is the constraint on everything. Most attempts to speed it up target the reviewers, which is why they fail.
Life Sciences Commercial practice
White papersMay 202611 min read
A practical sequence for cloud cost control
Cloud spend rarely grows because of one bad decision. It grows because nothing makes the cost of a decision visible at the moment it is taken.
Cloud & Infrastructure practice
Updated quarterly
Technology Radar
Where we place enterprise technologies today — what we would recommend adopting, what is scaling, what is worth a bounded pilot, and what we are simply watching.
Explore the radar →
Browse
By format
Articles
2 published
Executive guides
1 published
White papers
1 published
Webinars
0 published
Industry points of view
1 published
Case studies
Everything we have published
Browse insights
Filter
5 results
Articles7 min read
Why AI pilots stall before production
Most enterprise AI pilots work. Far fewer reach production, and the reason is almost never the model.
Data, Analytics & AI practice
Articles6 min read
Reducing testing time without reducing coverage
When a regression pass takes two days, the instinct is to automate more of it. The faster answer is usually to move it to a different layer.
Quality Engineering practice
Executive guides9 min read
Replatform, refactor or rebuild: deciding on evidence
Three legitimate answers to the same legacy problem, with very different costs. An executive guide to choosing between them on evidence rather than preference.
Product & Platform Engineering practice
Industry points of view6 min read
Modular content and the medical review constraint
In life sciences commercial operations, review is the constraint on everything. Most attempts to speed it up target the reviewers, which is why they fail.
Life Sciences Commercial practice
White papers11 min read
A practical sequence for cloud cost control
Cloud spend rarely grows because of one bad decision. It grows because nothing makes the cost of a decision visible at the moment it is taken.
Cloud & Infrastructure practice
Editorial standards
How we publish
- Every article is attributed to the practice that wrote it and the function that reviewed it. Named individual bylines are added as publication consent is recorded — we will not put a person on a page before they have agreed to it.
- Publication and last-updated dates are both visible when they differ.
- External claims carry a citation. Internal claims carry evidence or they do not run.
- Content is reviewed on a fixed cycle. Anything unreviewed past its interval is flagged rather than left to quietly age.