Five agents that work the register between closes.
AssetIQ's agents are deterministic where it matters and gated by human approval everywhere it counts. None of them can post an external, irreversible entry on their own.
How every agent is governed
- Draft, don't commit. Agents produce reversible artifacts a human reviews.
- Deterministic math. Depreciation, valuation, and matching are rule-based, not LLM-guessed.
- Confidence-aware routing. Low-confidence proposals always escalate to a human.
- Full provenance. Every run, tool call, and approval feeds the audit trail.
Agent
Reconciliation Agent
Floor-to-books reconciliation used to mean weeks of spreadsheet chasing and undetected ghost assets.
See how it works →AgentDepreciation & Compliance Agent
Dual-book depreciation — Companies Act Schedule II and IT Act §32 blocks — computed by hand, every close.
See how it works →AgentPhoto-to-Register Agent
New assets sit in a spreadsheet purgatory between procurement and capitalization.
See how it works →AgentGST / Transfer Agent
Inter-GSTIN asset transfers are a uniquely Indian tax trap — miss one and it's a ₹10,000 penalty per instance.
See how it works →AgentAudit Trail Agent
Assembling evidence for a statutory auditor takes weeks of pulling logs, screenshots, and sign-offs.
See how it works →Frequently asked questions
- Can an AI agent post an entry to my ledger without approval?
- No. Every agent in AssetIQ follows a draft-don't-commit model — agents can propose journal drafts, exceptions, and transfer orders, but a human with approval authority must review and post anything that reaches the ledger, GST portal, or e-way bill system.
- Do the agents do the depreciation math themselves?
- No. Depreciation math, GST valuation, and scan-to-asset matching are computed by deterministic tools, not the language model. The AI orchestrates, classifies ambiguous cases, and writes the explanation — it never performs the arithmetic that ends up in the books.
- What happens if an agent isn't confident about a result?
- Each agent attaches a confidence score to every proposal. Below a per-agent, per-materiality threshold, the item is routed to a human for mandatory review, even if policy would otherwise allow it to auto-commit as an internal draft.
- Is every agent action logged?
- Yes. Every agent run records its trigger, inputs, model version, every tool call and result, its confidence score, the human who approved it (if any), and the resulting records — feeding directly into the immutable, hash-chained audit trail.