The AI stack we own, not the one we resell.
iLeaf.ai is iLeaf Solutions' proprietary AI framework: retrieval grounded in a client's own material, a governor that enforces limits on spend, latency and permitted actions at runtime, and an audit trail recording the goal, tools called, reasoning and cost of every agent run. It runs underneath every product and engagement we deliver. The product site is at www.ileaf.ai, and Adaptive AI — the version that makes existing software agentic — is at ai-adaptive.ileaf.ai.
www.ileaf.aiThe framework's own site — go and see it.Ask Lia about what is ileaf.ai?What this includes
- Grounded retrieval
- Answers drawn from your own documents and knowledge graph rather than the open internet, with the source cited so a claim can be checked.
- The Governor Engine
- Runtime limits on spend, latency and permitted actions — an agent that meets something unfamiliar escalates to a person instead of guessing confidently.
- Full audit trail
- Every run records its goal, the tools it called, its reasoning, its cost and its outcome, so any decision can be reconstructed months later.
- Human-in-the-loop routing
- The confident majority handled automatically; low-confidence, high-consequence and adverse cases routed to a person with the context already assembled.
- Model-independent
- The framework is not a wrapper around one vendor's model, so the model underneath can change without the governance and retrieval changing with it.
- Private by default
- Client data is never used to train public models, and where it must stay inside your estate we have run the whole stack on a client's own hardware.
How we run it
Every phase ships something usable on its own, so you are never holding a half-finished system waiting on the next milestone.
Governance before autonomy
The gates come off last, not first. An agent shadows the existing process until its decisions can be compared against your team's on the same cases.
Built because we needed it
iLeaf.ai was not built to sell. It exists because running agents in healthcare, finance and legal work demanded answers to questions the model vendors do not answer.
One framework, every engagement
The same stack runs under LeadChirp, Legaltrack AI and bespoke client builds, so what is hardened on one project is hardened for all of them.
What we build it with
- Graph-RAG
- Vector retrieval
- MCP skill layer
- Runtime governance
- Agent observability
- On-premise deployment
Questions we get asked
What is iLeaf.ai?
iLeaf.ai is iLeaf Solutions' own AI framework — grounded retrieval, a runtime governor limiting spend, latency and permitted actions, and a full audit trail on every agent run. It sits underneath every product and client engagement the firm delivers. The product site is www.ileaf.ai.
Is iLeaf.ai a product I can buy?
No. iLeaf.ai is the framework underneath what iLeaf Solutions builds, not a licence sold on its own. Adaptive AI, at ai-adaptive.ileaf.ai, is the service delivered on top of it — an engagement that makes software a business already runs agentic, without rebuilding it.
How is this different from using ChatGPT or Claude directly?
A model answers questions; iLeaf.ai decides what an agent is allowed to do with one. The framework adds retrieval from your own material, enforced limits on cost and actions, escalation when confidence is low, and a record of every decision — none of which a model provides on its own.
Which model does iLeaf.ai run on?
It is deliberately model-independent. iLeaf.ai is not a wrapper around one vendor, so the underlying model can be changed — or run privately on a client's own hardware — without the retrieval, governance or audit layers changing with it.
Let’s talk about what is ileaf.ai?.
Tell us what you are running and what it needs to do next. We will tell you honestly whether we are the right team for it.