Notes from Josh + Amelia
Things we are seeing, testing and trying to make sense of as AI starts to show up in day-to-day work. The useful bits, rather than a stream of AI news.
A practical guide for owner-led and operations-led small businesses looking for sensible places to examine AI, starting with repeatable operational work rather than broad promises.
An AI workflow audit is a practical review of how work actually moves through a business, where time and judgement are being spent, and which improvements are worth testing first.
The best AI starting points are usually already visible in the work your team repeats, chases and rewrites. Here is a practical way to find them and choose a first project that is worth doing.
Microsoft Scout is an early signal of where AI agents are heading: away from demos and side panels, and into the systems where real work happens.
Strict security is not the enemy of AI adoption. Bland, over-controlled rollouts are. Here is how companies can give AI enough room to be useful without losing control.
Dashboards show what happened. KPI agents help teams decide what to do next, assign the work, and keep performance moving.
Dashboards still matter, but they make busy teams hunt for problems. AI agents can bring the right signal, context and action to the right person.
The companies that get real value from AI will not be the ones with the most tools. They will be the ones that integrate AI agents into the way work actually happens, while humans stay responsible for mission, standards and direction.
Claude Opus 4.7 matters less as a benchmark headline and more as a sign that AI tools are becoming more dependable, assignable, and useful in real business work.