back to work

ai operations platform for small trucking carriers

An MVP focused on explainable payment-risk and operational alerts for carriers working across disconnected systems.

contextearly-stage logistics product
timeline2026 — now
teamproduct, engineering, carrier operators
my roleproduct discovery, MVP definition, technical feasibility
statusMVP in progress
5+fragmented source types
MVPcurrent product stage
pilotnext validation step
0→1product scope

overview

The MVP links financial and operational evidence to explainable alerts. Current work is validating invoice-payment matching, revenue alerts, broker concentration, workflow audits, and recommended actions without claiming pilot impact prematurely.

the user — Small trucking carriers with limited software infrastructure.

the problem

Invoices, email, bank activity, factoring documents, and operational records rarely agree. Payment risk, revenue changes, and broker concentration can be discovered too late.

what i owned

  • mapped fragmented carrier workflows and source data
  • defined validated problems and current product assumptions
  • translated financial records into user-facing product states
  • shaped MVP requirements and feasibility questions
  • defined planned pilot measurements

evidence and decision sequence

01
observed signalinvoice, bank, factoring, and email evidence
product decisionnormalize records into linked workflow events
how it was checkedtest entity and payment matching accuracy
02
observed signallate invoice or unmatched transaction
product decisionshow an explainable alert, not an autonomous action
how it was checkedoperator rates usefulness and takes or rejects action
03
observed signalrepeated pilot use
product decisionexpand only the alerts that change decisions
how it was checkedresolution time, repeat use, and observed action
pilot gate

expand only after a narrow alert proves useful

1. match

Can the product link invoice and payment evidence accurately?

2. explain

Can an operator understand why the alert appeared?

3. act

Does the alert change a follow-up or financial decision?

4. repeat

Does the operator return to use the workflow again?

key decisions and trade-offs

decision

start with explainable alerts, not autonomous actions

alternativesBuild a broad chatbot, automate financial actions, or provide a static dashboard.

what we choseDetect a narrow set of high-value conditions and show the evidence behind each recommendation.

whyCarrier operators need to trust where an alert came from before delegating action.

trade-offThe MVP covers fewer workflows, but each has a clearer validation path.

decision

treat documents as evidence in a shared model

alternativesBuild separate tools for invoices, email, banking, and factoring.

what we choseNormalize records into linked entities and workflow events.

whyThe value comes from reconciling sources, not summarizing each one separately.

trade-offEntity resolution is harder, but supports more useful cross-source insights.

product interface

payment-risk alertsanitized portfolio reconstruction
source evidenceinvoice #1842
amount
$4,850
days outstanding
47
broker
northline freight
bank matchnot found
payment risk detectedbroker concentration: 43%

invoice is past the expected payment window and no matching bank transaction was found.

invoice parsedbank checkedrisk flagged

evidence and next steps

validated problem

Operational and financial evidence is fragmented across disconnected systems.

current assumption

Explainable payment and concentration alerts may create more value than a general assistant.

next experiment

Pilot invoice matching and measure accuracy, resolution time, alert usefulness, and repeat use.

measurement notes

Definitions shown so the claims can be discussed and challenged.

matching accuracy

Planned pilot measure: accepted invoice-to-payment matches divided by reviewed matches.

resolution time

Planned pilot measure: time from an alert to a verified operator outcome.

alert usefulness

Planned pilot measure: operator rating plus evidence that the alert changed an action.

reflection

The priority is proving that a narrow alert changes a real decision. Assumptions remain explicitly labeled until pilot evidence supports them.