Marco runs a twelve-person digital agency, and his margin problem is the same one every services firm has: every client engagement starts from scratch. Botmanor's multiplication machinery — templates, store packages, and workspace-per-client isolation — is built to break that pattern.
The arc starts with structure. Each client gets their own workspace, so data, knowledge bases, conversations, and billing are isolated by construction, not by discipline. The agency's delivery pipeline authenticates with machine credentials that go through the same enforced RBAC as human users — so installs and updates can be scripted instead of clicked.
Then Marco builds the asset: a "Client Support Bot" refined on the first engagement, captured as a store package. The package carries the app, its agents with their published versions, prompt templates, and the retrieval profile that made answers feel right.
Delivery becomes an install decision. Fast-follow clients get a linked install that tracks the agency's upstream improvements — harden the prompt once, every linked client can take the update. Clients with bespoke requirements get a forked copy the team customises freely, or a fork-after-install when a linked client outgrows the standard build. Per client, the team loads that client's documents into a knowledge base, validates retrieval with semantic search, and takes the bot live on the client's own Slack or Teams via channel connectors and routes.
Operating it stays centralised. Every agent run in every client workspace writes an execution record with token and cost metrics, so Marco can see per-client economics — which bots are heavily used, what each one costs to serve, and where a model swap would improve margin — before the monthly invoice goes out. One build, many deliveries, and the oversight to price it properly.
This story is a cross-workspace delivery practice, not a single in-app walkthrough: the interactive journey would need per-client workspace creation, machine-credential setup, and store installs spread across multiple surfaces, so it is presented here as a narrative pattern rather than a clickable "Do it yourself" section.
Ready to make this your story?
