Set the objective. Then try to break it.
Nash's homepage explains autonomic logistics in three steps. Here you can run them yourself, on a live city.
- Planned route
- Nash acted
- The world pushed back
1Set the objective
2Throw something at it
3Watch it hold
- Average wait
- 0min
- Route distance
- 0km
- Longest wait
- 0min
Average wait over the last minute. It spikes when you interfere, then settles.
- 0 delivered
- 0 disruptions absorbed
- 0 manual fixes
What I took from nash.ai, and what I changed
The homepage has a three-step flow: set the objective, Nash optimizes until it's met, outcomes compound. It is the core of the pitch, and today it is three blocks of text. The product video is still marked as coming soon.
The brand post says Nash would rather show systems working than claim they do. So I turned the explainer into something a visitor can operate. A buyer who closes a road and watches the routes heal has understood the product before reading a single feature.
Where AI did the work
I built this with Claude in one sitting. I gave it Nash's published brand rules as hard constraints, described the scenario, and directed the result the way I would direct a junior motion designer: what to keep, what to cut, what felt off-brand.
How it scales the brand
Change the vertical in the top corner of the map. Same engine, different vocabulary. The scenario is one small config object, so the same piece becomes a grocery hero loop, a pharmacy sales demo on a customer's own city, a booth screen, or a ten-second social clip. Each one comes out on-brand because the rules live in the template, not in a reviewer's memory.
The brand rules, written as constraints
- Chartreuse means Nash acted. It appears when a route is re-planned or a delivery is confirmed, then fades. It is never decoration.
- Disruption shows as aberration. Affected routes shake and split into colour channels, then snap back sharp once the plan holds.
- The world is a hex grid. Every route follows it, and closures take cells out of it.
- Motion leaves light trails. Couriers stay sharp while their path blurs behind them, like the long-exposure photography.
- Aurora behind everything. The background gradients are still resolving when the page loads.
- Grounded copy. The log names the friction: a closed road, a courier offline. Nothing here says "AI-powered".
The prompt template behind it
You are extending the Nash brand, "Equilibrium in Motion". Build a [format: hero loop, sales demo, booth screen, social clip] for [vertical]. Rules that never bend: 1. Chartreuse appears only after Nash has acted: a route confirmed, an exception cleared. Never use it as decoration. 2. Show disruption as aberration. The affected element shakes or splits into colour channels, then snaps back into sharp focus. 3. The world is a hex grid. Movement leaves light trails, and the Nash element stays sharp. 4. Backgrounds are slow aurora gradients that look like they are still resolving. 5. Copy is grounded. Name the friction (weather, traffic, a missed scan). Never write "AI-powered". Scenario: [disruption] hits [number] routes while the objective is [objective]. Sequence: objective set, disruption, re-plan, equilibrium regained. Vocabulary: call a stop [order / prescription / meal] and a courier [driver / courier / rider]. Output one self-contained HTML file. Draw on canvas, respect reduced-motion settings, and keep every number illustrative.