AI assistant
AI across every step of the trade.
Groq-powered LLaMA models work throughout the platform: drafting documents by chat, scoring and strengthening them, analyzing deals, passports, and listings for soundness and fraud risk, researching counterparties, and answering trade questions with memory and voice. Thirteen distinct tools, every one grounded in your actual data.
13
AI tools
LLaMA 3.3 70B
Plus a fast model
0-100
Soundness scoring
Deterministic
Repeatable analysis
Why it matters
Built to be believed.
Grounded, not generic
Every tool works from your real deal, document, or screening data, and is barred from inventing prices, parties, or sanctions verdicts.
Drafting to diligence
One assistant spans the whole workflow: write the document, score it, analyze the deal, and research the counterparty.
Repeatable analysis
Soundness scoring runs deterministically, so the same input yields the same score, not a different number each time.
Drafting
Build documents by conversation
Describe the deal in plain language and the assistant gathers every required field, normalizes quantities, currencies, incoterms, and dates, and hands you a complete draft. Per-field generation fills any single clause on demand.
- Conversational document creation across all 16 document types
- Grounded in your profile, with your company pre-filled correctly
- Per-field AI generation for individual clauses
- Never invents party names, amounts, or terms
AI document chat
Drafting500 kg gold, CIF Zürich, irrevocable LC at sight.
Fields filled from your words · never invented
Quality
Score and strengthen every document
An AI audit scores a document from 0 to 100 across completeness, language, legal protection, specificity, and compliance, then one click rewrites the weak sections while preserving every factual value.
- 0 to 100 quality score across five dimensions
- Specific, actionable suggestions per dimension
- One-click apply fixes rewrites weak sections only
- Preserves all party names, amounts, dates, and terms
Quality score
3 fixes86
of 100
One click rewrites weak sections, values preserved
Risk
Soundness and fraud analysis
AI analyzes a deal, a provenance passport, or a marketplace listing for soundness and fraud risk, scoring completeness, consistency, counterparties, documentation, and classic fraud signals, with red flags and concrete suggestions.
- 0 to 100 soundness score with a low, medium, or high risk level
- Works on deals, provenance passports, and marketplace listings
- Surfaces specific red flags and fraud signals
- Runs deterministically for repeatable results
Soundness analysis
Low risk78
Completeness · consistency · counterparties
Diligence
Research, benchmarks, and clauses
Run AI due diligence that combines web research and recent news signal into a structured company summary, get market context for your deal terms, clause recommendations, and a plain-language narrative on screening results.
- Due diligence with web research, news signal, and a structured summary
- Market benchmark context on your deal terms
- Clause recommendations tailored to commodity and route
- Plain-language risk narratives on screening results
AI due diligence
Research aidMeridian Commodities Ltd
News signal: 5 recent articles, mixed tone
Not an authoritative sanctions verdict
Assistant
A trade assistant with memory and voice
Ask about Incoterms, payment methods, document drafting, or deal terms in a multi-turn conversation that remembers context across sessions, with voice input when you would rather speak than type.
- Multi-turn conversation with persistent history
- Voice input on the assistant and document chat
- Trade-domain expertise on Incoterms, payments, and procedures
- Quota-managed per plan, from the Free tier to unlimited on Enterprise
AI assistant
Memory + voiceCIF includes freight and insurance to the named port. CIP extends that to any transport mode.
How it works
From start to proof.
Describe what you need
Draft a document by chat, or open the assistant and ask a trade question by text or voice.
Let AI do the work
Generate fields, score quality, analyze a deal or passport, or research a counterparty.
Review grounded output
Every result is built from your real data, with suggestions, red flags, and a clear rationale.
Act with confidence
Apply fixes, address red flags, and move the deal forward with the evidence in hand.
FAQ
Common questions.
Which AI tools are included?
Thirteen, spanning conversational and per-field document drafting, quality scoring and one-click fixes, soundness and fraud analysis on deals, passports, and listings, due diligence research, market benchmarks, clause recommendations, risk narratives, and a chat assistant with memory and voice.
Does the AI make up prices or sanctions verdicts?
No. The models are explicitly barred from fabricating specific market prices or statistics, and AI commentary on screening is never an authoritative sanctions verdict. The real screening verdict comes from the compliance engine.
Is the analysis consistent?
Soundness scoring runs deterministically, so identical inputs produce identical scores rather than a different number on each run.
Are there usage limits?
AI is quota-managed per plan, scaling from a small monthly allowance on the Free tier through to unlimited on Enterprise. Heavier operations such as full-document rewrites and due diligence cost slightly more.