Tune your agents
without writing code.
The Agent Workshop organizes your entire AI configuration around six agent personas — Sales, Support, Troubleshooter, After-Hours, Orchestrator, and QA Reviewer. Edit personality prompts, toggle capabilities, adjust tuning parameters, and test everything with the built-in Simulation Studio. Sierra charges $150K+ for their Agent Studio. This ships with every Vectrant plan.
Six agents, one workshop.
Each agent has its own personality prompts, capability toggles, and tuning knobs. Change one without affecting the others.
Sales Agent
Product recommendations, bundle suggestions, financing guidance, inventory checks, and commercial intent detection. Four prompt sections and seven capability toggles.
Support Agent
Knowledge base queries, multilingual support, frustration detection, and cross-encoder reranking. Confidence thresholds let you control when the AI answers vs. admits uncertainty.
Troubleshooter
Guided DIY fix pipeline, web search for repair guides, and automatic service claim creation when troubleshooting fails. Three toggles control the full flow.
After-Hours Agent
Collects contact info for morning follow-up when the store is closed. Configurable business hours with Eastern timezone awareness.
Orchestrator
Routes every query to the correct pipeline — products, knowledge base, orders, troubleshooting, or shopping flows. Confidence threshold and token limit are tunable.
QA Reviewer
Reviews overnight conversations every morning and generates quality reports with per-conversation evaluations. Enable or disable with a single toggle.
Simulation Studio: test before you ship.
Run realistic multi-turn customer conversations against your live or draft configuration — directly in the browser. No staging environment, no guesswork.
One-click simulation
Select a scenario from the built-in library or create your own. Hit Run and watch a realistic customer-agent conversation stream in real time, token by token.
Draft mode testing
Edit a prompt, save it as a draft, and run a simulation against the draft config — without publishing. See how changes affect behavior before they go live.
Full conversation replay
Every simulation run creates a real conversation session. View it in the Conversations page with full metadata, timing, and turn-by-turn analysis.
Agent Workshop
Personality prompt
PublishedCapabilities
7 activeSimulation Studio
Draft configGlobal rules that apply everywhere.
Accuracy rules, brevity rules, and custom instructions apply across all agents. Edit them once, and every persona inherits the change.
Accuracy & brevity rules
Cross-cutting prompt sections that enforce factual accuracy and response length across every pipeline. Published or drafted independently of individual agents.
Custom instructions
Add your own behavioral rules that don't fit a specific agent — company policies, holiday messaging, seasonal promotions. Create, edit, and publish without touching code.
Agent Workshop, running on your data.
30-minute walkthrough on your real catalog. No slideware, no demo data.