Why PhantomOps, why now
A sovereign, Arabic-native agentic AI platform for BFSI in the GCC — proven inside our own business, with regular updates published here as we build.
We are our own first customer
Sally is not a prototype. She runs today on a LangGraph state machine, monitors a live mailbox, and turns a single manager email into finished, sourced work — across four LLM providers at once.
Source: PhantomOps / Sally internal build — LangGraph · Outlook & Gmail · Groq, Gemini, Claude, OpenAI.
From internal capability to regional product
Foundation & efficiency
- Sally, Mark & Critic agents live
- Document RAG & sovereign corporate brain
- VM + email interface, persistent memory
- Multi-provider LLM gateway
Multi-modal & fleet
- Noor Arabic agent for the GCC
- PI Guard, Magic Reader & Arabic NLP Processor productized
- Explainer video & richer media
- Autonomous bid pricing & clause analysis
- Team of four running a 50+ agent fleet
PhantomOps.ae product
- Multi-tenant, cloud-scalable platform
- Public free tier for enthusiasts
- BFSI managed-service offering
- Regular investor updates on this site
Build log
We'll keep adding to this log as PhantomOps progresses — the most direct way for investors to track real progress rather than a static pitch.
Sally goes autonomous
Dynamic agent configuration, user-controlled data sharing, and self-learning retrieval for files shared by managers as templates. Isolated code, data and application runtime to keep future releases from overwriting agent state.
Sally live: email-to-action, multi-model
Sally began running in her own VM, processing manager requests by email across multiple external LLMs with no human intervention required for the core loop, and handling document attachments (PPT, Word, Excel, PDF) end to end.
PhantomOps kicked off inside EAI Systems
A team of four engineers began building PhantomOps to automate EAI's own Sales and Marketing functions — the internal proof-of-concept behind the platform now being brought to BFSI customers.