Main Dashboard
Account Chat Room
Human-Agent-Collaboration
Agent-Agent Collaboration

SalesHQ — Control-Tower for Humans + AI Agents

Inspiration 🚀

Modern sales teams are field-testing “digital teammates” (LLM-powered bots that prospect, research, write outreach, and schedule meetings).
But today those bots live in separate dashboards, noisy Slack channels, and siloed CRMs.
We kept hearing the same complaints from sales leaders:

“I’m drowning in bot pings.”
“I can’t see why the AI decided to do that.”
“Every new agent means another tab.”

We thought: what if a sales org had an air-traffic-control tower that showed deals, bots, and humans in one place—so reps steer while agents hustle in the background?

What We Built 🛠️

SalesHQ is that tower:

Unified Dashboard — Kanban cards for every deal with live agent avatars.
Account Rooms — Slack-style threads where bots drop drafts, humans approve inline.
Silent Inter-Agent Comms — Bots whisper via the Google A-2-A protocol; no human spam.
Extras — ResearcherBot pulls fresh facts with Apify

All of it runs end-to-end in TypeScript.

How We Built It ⚙️

Layer	Tooling
AI Agents	AgentKit workers (Prospector, Researcher, SDR, Scheduler)
Inter-Agent	Google A-2-A protocol tasks / results
Frontend	Vite (dark, neon-teal theme)
External APIs	Apify Website Crawler

What We Learned 📚

Agent orchestration is UX, not just code. Users love “silent” coordination far more than raw chat dumps.
A-2-A Protocol makes multi-agent pipelines trivial once the JSON schemas are right.
Apify MCP saves brain cycles

Challenges We Faced 🧗‍♂️

Noise vs. Insight — Tuning when bots post versus whisper took multiple UX iterations.
A-2-A Implementation

What’s Next 🌱

Fine-grained permissions so an SDR-bot can’t email Tier-1 accounts without human sign-off.
Open-source agent templates so any dev team can plug their own bots into SalesHQ.

Built in 8 caffeine-fueled hours by Bayram & Merdan.