AI & ML interests

None defined yet.

zchmael 
posted an update 10 months ago
view post
Post
225
BIG news from Averi: We just launched The AI Marketing Workspace – a complete reimagining of how marketing teams work in the AI era.

This isn't another AI tool you add to the stack.
It's the workspace that replaces the stack.

How marketing flows in Averi:

Plan → Strategy with AI that knows your brand and was trained by some of the best marketers on the planet. Activate experts for pressure-testing.

Create → /create mode drafts in your voice, and gives you the controls to edit right in platform. Bring in specialists when you need polish.

Execute → Deploy with full context. Experts see everything—no re-briefing.

Scale → Library captures what worked. Every project makes the next one easier.

All in one workspace. AI + human expertise at every phase. Context that never dies.

What makes it different:
→ Everything in one place – AI assistant, expert network, content library, execution tools

→ Shared context across all phases – Your strategy informs creation, creation informs execution, execution informs future strategy

→ /create mode's three-phase flow – Discuss → Draft → Edit takes you from blank page to deployment-ready

→ Seamless AI + human collaboration – Activate experts mid-conversation, edit together in real-time

→ Built like a workspace – Multiple tabs, organized folders, persistent memory

At its core is /create mode: where AI drafts brand-aligned campaigns, blogs, emails, ads, and social posts—then you and your experts refine them in one canvas.

No more copying between tools. No more losing momentum. No more re-briefing freelancers on your brand.

This is how marketing teams should work. Unified. Intelligent. Human.

👇 See the workspace in action: https://www.averi.ai/
zchmael 
posted an update about 1 year ago
view post
Post
256
Synapse: Multi-Model Routing Architecture for Marketing Tasks

Hey HF community! We just released Synapse, a system that orchestrates multiple models + humans for marketing workflows.

The Problem: Generic LLMs struggle with brand consistency, don't know when to think deep vs. fast, and can't decide when human expertise is needed.

Our Approach:
AGM-2 (13B model trained on 2M marketing docs)
Frontier models (GPT-4/Claude) for generalization
Human experts activated programmatically
Adaptive reasoning that scales cognitive effort automatically

Architecture: 5 specialized "cortices" handle different functions:

Brief parsing → Strategic planning → Creative generation → Performance optimization → Human escalation

Results:
Consistent brand voice, 2-3 week workflows down to 72 hours, measurable improvement over ChatGPT-based systems.

Questions for the community:
Thoughts on multi-model routing in production?
How do you handle human-in-the-loop at scale?
Anyone working on similar domain-specific architectures?

Built this to solve real marketing execution problems.

Would love feedback from researchers working on routing, domain adaptation, or human-AI collaboration.

Links: https://www.averi.ai/blog/averi-launches-synapse-a-new-ai-system-architecture-powering-agm-2-and-human-ai-collaboration-at-scale