Thinking & sub-agents Coming soon
For a hard task, Mapsly AI thinks longer before answering: it plans, weighs its options and returns one result. When optimal, it spawns sub-agents on appropriate models and reasoning effort, splitting the work into smaller tasks, reducing context load and staying sharp longer.
How it works
- 1Set the effortOn Enterprise, an admin sets the reasoning effort per conversation: quick for everyday questions, deep for planning a week.
- 2ThinkIt works the task in passes. When optimal, it spawns sub-agents on the right models and effort, in parallel.
- 3PublishThe result arrives as one plan with its numbers. You publish it; nothing changes without you.
reasoning effort: high passes: plan → options → compare sub-agents: coming soon tool calls: 24 model: frontier tier result: 5 days · 6 reps · 31 stops driving 22 h (−4 h)
- Reads
- Whatever the task needs: calendars, routes, accounts. Deep thinking reads more.
- Writes
- Only what you approve at the end. Thinking itself writes nothing.
- Needs from you
- A reasoning effort per conversation, set by an admin on Enterprise.
- Runs on
- Enterprise, on any model or your key. Essential and Pro: fixed effort.
Mapsly AI is included in every plan and is in private beta: contact us to join the waitlist and get early access. Each plan carries a monthly AI allowance measured in AI credits (the amount in brackets is what those credits currently cost), pooled across your account and reset every month. Beyond it, usage draws on your prepaid AI credit balance at the current rate card.
See Mapsly AI on your own pipeline
A personalized demo with a Mapsly solution engineer, on your data. Configuration is included in every plan.