One platform for supplier intelligence, sourcing, and trade compliance
These are not three tools with a shared login. Supplier data feeds sourcing, sourcing feeds classification and landed cost, and all of it lands on one record with one audit trail. That connection is the product.
Why it is one platform and not three
Each capability produces something the next one needs. Run them in separate tools and you spend your week reconciling them, which is the work Infis is trying to remove.
- ERP supplier master
- Quality records
- Regulatory and customs feeds
- Requisitions and BOMs
- Qualified alternates
- Your RFQ templates
- Part specs and drawings
- HTS schedule
- Section 232 scope
It watches, rather than waiting to be asked
The difference between this and a risk report is not the data, it is the clock. A report tells you what was true when somebody commissioned it. Supplier Monitor runs against your whole base continuously and tells you when something changes, the day it changes.
- Supplier master and site level records from ERP
- Deviations, CAPAs, audit findings, and certificate expiry from your quality system
- Regulatory actions, inspection outcomes, and import alerts
- Customs filings, financial disclosures, ownership and sanctions data
- Resolves the same supplier across systems, including sites your ERP holds as separate vendors
- Maps Tier 2 relationships from customs and filing data, so you see behind your direct supplier
- Scores compliance, financial health, delivery, concentration, and geopolitical exposure separately
- Re-scores on every new signal and routes anything past your threshold to the category owner
- A live score per supplier, decomposed into the factors that produced it
- An alert with the source document attached, not a summary of one
- Exposure in dollars, joined to the materials and open POs behind it
- A Form 483 published on Tuesday is scored and routed on Tuesday, with the filing linked, rather than surfacing in the next quarterly review.
- Sole source exposure is expressed in dollars against real open POs, so the qualification budget argument stops being a matter of opinion.
- Sites your ERP holds as three separate vendor numbers resolve to one supplier, which is usually where the concentration was hiding.
Built for the several hundred line package, not the three line one
Sourcing tools tend to be demoed on a tidy example. The job that actually eats a buyer's month is a long BOM going out to a supplier set that will not all reply, coming back in four different formats. That is the case this is built around.
Every line, against suppliers qualified to make it
The agent parses the requisition or BOM, normalizes part numbers against your material master, and matches each line to suppliers on capability, qualification status, lead time, and past performance. Lines it cannot confidently match are separated out rather than guessed at.
You review the supplier list before anything is sent
Add, remove, and re-rank. Nothing has left Infis at this point. This gate stays closed by default for as long as you want it to.
Drafted from your template, sent from your domain
The package is split by supplier scope so each one receives only the lines they were asked to quote. Follow ups go out on your cadence, and the agent tracks who has opened, replied, and gone quiet.
Four formats become one comparison
Currency converted, unit of measure reconciled, incoterms made like for like, and tiered price breaks resolved at your actual volumes. Non responders show as no bid rather than disappearing.
The rows a careful reader would have caught
Price outliers against the rest of the field, lines down to a single qualified source, expired certifications, and anything the agent scored low confidence. These are raised, not left in the sheet to be noticed.
The award is yours
The agent did the assembly. The decision, and the relationship consequences that come with it, stay with the buyer.
- A 340 line package goes out in an afternoon. The buyer spends their time on the supplier list at the front and the award at the end.
- Quotes come back comparable. Currency, unit of measure, and incoterm are reconciled before you see them, so nobody rebuilds a sheet by hand.
- Exceptions are found rather than noticed. Price outliers, lines down to a single qualified source, and expired certifications are flagged, not left for whoever reads carefully.
It classifies independently, then shows you the rule
Most classification assistants confirm whatever heading you put in front of them, which is worse than useless when the heading is wrong. The Classification Agent runs its own GRI analysis first and tells you when it disagrees, citing the rule and the source text that took it somewhere else.
- Works the General Rules of Interpretation in order, against the current schedule
- Quotes the source text it relied on rather than paraphrasing it
- Scores its own confidence, and routes anything below your threshold to a human
- Files the full reasoning against the part, so the record survives staff turnover
- Flags parts and materials that fall inside the pharmaceutical proclamation scope
- Models duty at the applied 232 rate rather than the MFN rate
- Recalculates landed cost across every affected material when scope or rates move
- The classifier reaches its own heading. If it disagrees with the one you proposed, it says so and shows the rule that took it elsewhere.
- Every answer carries the GRI applied, the source text quoted, and a confidence score, which is what makes it defensible when a broker or an auditor pushes back.
- Section 232 pharma scope is modelled at the applied rate, so landed cost reflects the duty you will actually pay.
The part nobody asks about until the audit
The pieces share data, not just a login
The score that flagged a supplier, the RFQ that went out because of it, the alternates that came back, and the classification on the part you switched to are all the same record. You can read the chain forwards or backwards.
Every entry keeps its source and timestamp
Scores cite the filing. Classifications cite the rule. Agent actions keep their trigger, the records read, and the person who approved them. Nothing in the trail is a summary of something you can no longer see.
Exportable, because reviewers want their own copy
The full action log exports as CSV or comes out over the API. When quality or an auditor asks why a supplier changed, the answer is a query, not a reconstruction across three tools and two people who have since left.
What buyers ask on the first call
No. Connectors are read only by default and scoped to the objects the model needs. Scores, alerts, and classifications live in Infis and are exportable, or pushed to Slack, Teams, and email. If you later want write back on a specific object, that is a deliberate change with its own scoped credential, not the default.
Start with the supply base you already have
Send a supplier list or an open RFQ. We will run it through Infis and show you the scored result, including the exposure your current process is not surfacing.