Institutional knowledge
Preserve business context, project history, decisions and the reasoning behind them so knowledge does not disappear with conversations or people.
Client work · Pune real estate
Building an enterprise that can remember, understand and act.
For an established real estate company, AVIS is building a Hermes-based second brain and enterprise management system—connected to an AWS-based curated information foundation where AI can analyse the business with richer context and produce decision-ready insight.
The client is intentionally anonymised. This page describes the active programme direction; individual capabilities should not be read as production-complete unless explicitly stated.
The connected system
One operational memory. One governed intelligence foundation.
The programme joins daily management, institutional knowledge and analytics instead of treating them as separate applications.
01Capture
Bring decisions, commitments, projects, operating knowledge and relevant business evidence into a durable enterprise memory.
02Curate
Transform source information into structured, contextual and governed knowledge on AWS.
03Analyse
Allow ANA and specialist agents to examine connected evidence, identify patterns and explain what matters.
04Act
Convert insight into decisions, ownership, follow-through and measurable operating improvement.
01 · HermesMEMORY MANAGEMENT AGENTS
A second brain that also helps run the enterprise.
Hermes is being designed as more than a document repository. It provides the working memory and management layer through which knowledge, priorities, decisions, initiatives and action remain connected.
Enterprise coordination
Connect goals, projects, commitments, owners and reviews so the organisation can move from aspiration to accountable execution.
Operational orchestration
Give authorised agents a governed context from which they can research, analyse, prepare work and support follow-through.
02 · AWS analytics foundationCONTROL CONTEXT TRUST
Curated information before artificial intelligence.
The analytics foundation is being built on AWS to organise relevant enterprise information, preserve meaning and make the evidence usable by analytical agents. AI quality depends less on how much content is stored than on whether the right context, definitions, provenance and relationships survive.
Known sources and ownership
Information can be associated with its source, business meaning, owner, review state and permitted use.
Business-ready knowledge
Metrics, projects, customers, processes, policies and decisions can be described as concepts—not anonymous text fragments.
Traceable analytical evidence
Material insights can retain a path back to the information, assumptions and definitions from which they were produced.
Google Open Knowledge Format
Why curated OKF knowledge is stronger than a conventional RAG-only approach.
Traditional vector RAG often divides source documents into chunks and retrieves passages that appear semantically similar to a question. That is useful for discovery, but similarity alone does not guarantee that a passage is authoritative, current, correctly defined or connected to the right business object.
OKF represents knowledge as human-readable Markdown concepts with machine-readable YAML metadata. For enterprise analytics, that makes meaning and governance first-class rather than leaving them implicit inside retrieved text.
Meaning survives retrieval
Concepts preserve business definitions, descriptions, tags and links instead of relying only on arbitrary document chunks.
Provenance and trust are explicit
Sources, verification, trust signals, freshness, lifecycle and attestation can travel with the knowledge an agent uses.
Human and agent readable
People can inspect the same Markdown knowledge that agents consume, improving review, correction and accountability.
Portable and versionable
The knowledge is vendor-neutral, diffable in Git and exchangeable across tools without requiring a proprietary runtime.
Continuously maintainable
Agents and people can improve the same corpus over time while retaining history and identifying what is current.
Retrieval remains available
OKF is a knowledge format, not a replacement for every retrieval technique. Search or RAG can still be used over the curated corpus where scale requires it.
03 · ANA Intelligence Studio
A dedicated analytical organisation—not another dashboard.
ANA leads four specialist agents through the complete intelligence cycle: DARA establishes trusted data, KIRA builds verified knowledge, PRIA analyses and anticipates, and ACTA converts approved insight into accountable action.
DARA · Data
Acquisition, reconciliation and analytical readiness.
KIRA · Knowledge
OKF curation, research, provenance and verification.
PRIA · Insight
Predictive reasoning, patterns, scenarios, risk and opportunity.
ACTA · Action
Ownership, follow-through, escalation and verified closure.
The outcome
From scattered information to an enterprise that learns.
Hermes preserves and coordinates the operating context. AWS provides the curated analytical foundation. OKF makes knowledge portable and trustworthy. ANA and her agents convert that evidence into understanding and action.
Discuss an enterprise intelligence system →