I'M RED
01 · Overview AI support, simple controls
I'M RED is Red Acre's AI knowledge base and chatbot for broker support: an admin panel where support teams train the bot, and a customer-facing chat widget that answers in their brand.
02 · The challenge Built for support staff, not engineers
Support teams at trading firms drown in repeated questions, and the tools to train AI on answers are usually built for engineers. The brief: make training, content and theming manageable by support staff with minimal onboarding.
03 · The solution Complex AI, simple controls
Every screen keeps the same rule. Guided intelligence flows walk staff through training the bot, agent profiles give the AI its personalities, article and content management feeds the knowledge base, and invite flows onboard the whole team.
04 · The admin panel What is drifting
on the right
- Welcome & onboarding: the first run, free trial setup, sign up, invites and user management.
- Conversations: the live queue and the detail drawer agents actually work from.
- Guided intelligence: the training flows, the rule builder and its drawers.
- Knowledge base: collections, articles, the content editor and the page scraper.
- Agent profiles: each personality and the knowledge it is allowed to reach.
- Automations & settings: workflow rules, chatbot integration and widget theming.
05 · Conclusion Support runs it,
not engineering
Every AI support tool on the market assumed an engineer would operate it, which handed the cost straight back to the people it was supposed to relieve. Red Acre needed broker support teams to stop answering the same question by hand, and needed those teams to run the AI themselves. Buying the capability was never the hard part.
So I designed for the operator rather than the model. One rule held across every screen: complex AI, simple controls. Guided flows walk a support lead through training the bot, agent profiles turn model behaviour into a named personality with a defined scope, and the knowledge base is edited like a CMS instead of configured like a pipeline. Nothing asks the user to understand how the AI works in order to change what it does.
Repeat questions resolve in the widget instead of the queue. Everything behind it is owned by the support team: training flows, articles, agent profiles, widget theming. What used to be a ticket for engineering is now a dropdown for a support lead.
06 · Live The chatbot widget
The customer-facing side: a themable chat widget that cycles below through home, conversation, collections and categories, switching between light and dark mode exactly like the shipped product.
