Why We're Building Around the Workflow Problem in Trade Compliance
One thing we've noticed while looking at global trade compliance is that the problem isn't always a lack of information.
There is plenty of information.
The harder problem is connecting the information.
A company might have its product catalog in one system, HS classifications in spreadsheets, tariff data somewhere else, landed-cost calculations with finance, and regulatory updates coming through emails or separate databases.
Each piece works.
The workflow doesn't.
The problem becomes obvious at scale
Imagine an importer managing 2,000 SKUs across several suppliers and countries.
A tariff or regulatory change happens.
Now someone has to figure out:
Which products are affected?
Are their classifications still appropriate?
Does the tariff treatment change?
What happens to landed cost?
Which upcoming shipments are affected?
Does procurement need to contact suppliers?
Does finance need to update its assumptions?
None of these questions is particularly unusual.
The painful part is connecting them.
That's the product problem we're interested in.
We're thinking about compliance as a workflow
Instead of treating classification, tariffs, landed cost, and regulatory monitoring as separate activities, the model we're working toward looks more like:
Product Data → Classification → Tariffs → Landed Cost → Regulatory Monitoring → Impact → Action
The important part isn't the diagram itself.
It's what happens when information moves through it.
A change in regulatory information should ideally lead to a question about affected products, not just another notification in someone's inbox.
A classification change shouldn't exist independently from the tariff and cost implications that follow it.
And a compliance alert shouldn't stop at “something changed.”
It should help the right person understand what might need attention.
We wrote a deeper breakdown of this approach here: how a connected trade compliance workflow can work.
Where AI fits
We're also interested in where AI genuinely helps rather than where it is simply added because “AI” sounds good.
Trade compliance involves a lot of unstructured information.
Regulatory notices can be long. Product descriptions can be inconsistent. Trade measures can involve multiple conditions.
AI can potentially help with:
Extracting relevant information
Summarizing regulatory changes
Identifying entities and products
Matching changes against structured product data
Prioritizing potentially relevant alerts
But we don't think AI should blindly make the final compliance decision.
A better model is:
AI finds and organizes → structured systems connect → compliance experts validate → business teams act
That keeps humans involved where judgment actually matters.
The product challenge
The interesting engineering/product problem isn't just building another database of regulations.
It's building the connections between:
Products
Classification
Trade rules
Costs
Regulatory changes
People
Actions
That's where a compliance platform can potentially create leverage.
At "Borderline Genius Inc.", we're working on this broader problem through our trade compliance technology ecosystem, including Genius Workspace.
The long-term idea is simple:
Less searching. Less spreadsheet reconciliation. More useful compliance intelligence.
If you're building in B2B SaaS, compliance tech, supply-chain software, or another domain where the real problem is fragmented workflows, I'd be interested in how you've approached the same problem.
