Documentation · Solomon Agentic v0

01 Getting started 2-minute setup

Before testing with the Day 1 → Day 14 CSVs, reset to an empty state so the demo narrative starts from zero.

1Click Import data on the topbar.
2Choose Reset to empty and confirm in the prompt.
3The dashboard reloads with all counters at zero — KPIs at ₱0, no findings, no POs, no inventory, no suppliers shown.
4The topbar should now say Empty · Solomon Agentic v0 · waiting for Day 1 import.

Required before testing the 14-day pack. Skipping the reset means existing data mixes with imports and the Day-X counter will not be accurate.

02 Pilot demo data · 14-day import pack 14 CSVs

Each file simulates one day of operational events: purchase orders, deliveries, inventory movements, and route dispatches.

Download them in order, upload via the dashboard's Import data button, and Solomon AI will surface fresh findings on every import. Imports are additive — findings from earlier days persist as you upload more.

Day 1

day-01-operations.csv

Baseline calm

Routine restock and on-time deliveries. No risk signals yet.

Download CSV

Day 2

day-02-operations.csv

Baseline calm

Second day of normal feed. Production line A consumes PCBs.

Download CSV

Day 3

day-03-operations.csv

Slight PO uptick

Bigger resin lot ordered. Inventory rotates normally.

Download CSV

Day 4

day-04-operations.csv

First late PO on SUP-002

PO-2026-0146 slips 3 days. First weak signal hits the supplier feed.

Download CSV

Day 5

day-05-operations.csv

Sensor cover drops

SUP-006 also slipping. SKU-1006 days-of-cover drops to 11d.

Download CSV

Day 6

day-06-operations.csv

Critical Stock fires (SKU-1002)

PCBs cross the reorder threshold. SUP-002 5 days late.

Download CSV

Day 7

day-07-operations.csv

SUP-006 reliability slides

On-time rate at 64%. SKU-1009 dormant 95 days.

Download CSV

Day 8

day-08-operations.csv

Supplier concentration risk

SUP-002 single-source on 26% of electronics spend.

Download CSV

Day 9

day-09-operations.csv

Dead inventory surfaces

SKU-1011 discontinued cable reel 130 days dormant.

Download CSV

Day 10

day-10-operations.csv

High-severity supplier finding

PCB PO finally arrives — 11 days late. SUP-002 written up.

Download CSV

Day 11

day-11-operations.csv

Trapped capital escalates

Marketing SKUs accumulate. ₱5.2M trapped across dormant lines.

Download CSV

Day 12

day-12-operations.csv

Multi-agent convergence

SUP-002 fourth late PO + SKU-1006 cover at 3 days.

Download CSV

Day 13

day-13-operations.csv

Route variance spike

Manila→Cebu cost 22% above plan. SKU-1011 liquidation rec.

Download CSV

Day 14

day-14-operations.csv

Full risk picture

3 high-sev findings open across all 3 agents. Dual sourcing approved.

Download CSV

Click Run all agents in the topbar only if you want a clean re-detection. That action is destructive and replaces all current findings with a fresh snapshot.

03 CSV schema · all 4 systems IMS · WMS · TMS · FAMS

Each row represents one operational event. Solomon classifies the row by event_type and writes to the right system — IMS, WMS, TMS, or FAMS.

Event types and which system they populate:

event_type	System	What it does
`purchase_order`	IMS	Inserts a new PO with supplier, SKU, qty, value, and date
`po_delivery`	IMS + WMS	Updates PO status (received/delayed). On `received`, also bumps on-hand inventory at the warehouse
`inventory_movement`	WMS	Adjusts `on_hand` at a warehouse: `received` = +qty, `issued` = −qty, `no_movement` = touches dormancy timer
`route_dispatched`	TMS	Inserts a route dispatch (origin → destination), with planned vs. actual cost and variance %
`asset_service`	FAMS	Tracks asset maintenance state (`due` / `overdue` / `serviced`) per warehouse
`agent_finding`	Ignored	CSV-provided findings are skipped. Solomon AI generates findings dynamically from current state.

Column reference:

Column	Used for	Example
`event_type`	One of the 6 types listed above	`purchase_order`
`event_date`	ISO date — drives the Day-X counter and dormancy calculations	`2026-05-01`
`record_code`	PO number / route code / asset code (e.g., `RT-MNL-CEB-01`, `FL-MNL-01`)	`PO-2026-0150`
`sku_code`	References the SKU catalog (auto-created on first sight)	`SKU-1002`
`supplier_code`	References the supplier catalog (auto-created on first sight)	`SUP-002`
`warehouse_code`	References the warehouse catalog (auto-created on first sight)	`WH-MNL`
`quantity`	Units for the event (blank when N/A)	`500`
`value_php`	Peso value of the event (actual cost for route_dispatched)	`925000`
`status`	`ordered` / `delayed` / `received` / `issued` / `no_movement` / `dispatched` / `due` / `overdue`	`delayed`
`note`	Free-text operational note. For overdue assets, can include "`N days overdue`" to set the actual lateness	PCB shipment 3 days late

Route & asset code patterns:

RT-{ORIGIN}-{DESTINATION}-{SEQ} for routes (e.g., RT-MNL-CEB-01 = Manila → Cebu, sequence 01). Solomon parses the origin/destination from the code automatically.
{TYPE}-{WAREHOUSE}-{SEQ} for assets: FL=Forklift, RK=Rack, CV=Conveyor, GS=Generator, HV=HVAC. Example: FL-MNL-01 = Forklift #01 at WH-MNL.

CSV findings are ignored. Rows with event_type=agent_finding are intentionally skipped. Findings come from Solomon AI's live analysis of the operational state — never from the CSV.

04 How Solomon AI works 3 agents

Three agents continuously watch your data layer.

Supplier Exposure

Concentration, lead-time drift, single-source dependencies, on-time-rate slippage.

Critical Stock

SKUs at or below safety/reorder with active velocity, days-of-cover analysis.

Dead Inventory

SKUs with no outbound movement for 90+ days; trapped capital quantified.

After every import, Solomon AI analyzes the current state and emits structured findings via the Solomon AI model (or rules fallback if the API key isn't configured). Findings are additive — old ones persist so the demo accumulates risk signals as you upload Day 1 → Day 14.

Manual Run all agents / Re-run agent buttons do a destructive refresh, replacing existing findings with a clean snapshot of current data.

05 AI features across the app 4 touchpoints

Today's AI Briefing

Auto-generated executive summary on the dashboard — current risk level, biggest issue, and what to act on this week.

AI Analysis card per agent

3-5 line analysis specific to each agent, with named SKU/supplier references.

Explain (AI) per finding

Inline deep-dive on why the finding fired and what to do next, accessible from the button on each finding row.

Solomon Chat

Open-ended Q&A grounded in live data. History persists across page reloads and auto-clears after 1 hour. Erase icon clears it manually.

Each AI card shows a small pill: ⚡ Solomon AI when the model is responding, Rules on the deterministic fallback.

06 Reset & import behavior What changes when

Quick reference for which tables get touched by each action.

Action	Suppliers / SKUs / Warehouses	POs / Inventory	Findings
Reset to empty	Wiped	Wiped	Wiped
Import CSV file	Auto-create on first reference	Updated / appended	Additive (no wipe)
Run all agents button	Untouched	Untouched	Wiped & re-detected
Acknowledge / Snooze a finding	Untouched	Untouched	Status updated only

07 API integration · you provide the endpoint, we build the connector Production flow

Manual CSV uploads are for the pilot demo only. In production, we integrate Solomon with your existing system via API. You give us the endpoint — our developers handle the integration on the Solomon side.

The flow: You expose (or point us to) the read-only API endpoint of your ERP/WMS — for example, SAP S/4HANA's OData service, NetSuite's REST Web Service, or your custom REST endpoint. We consume that endpoint from Solomon and pull data on a schedule.

Systems we've integrated with:

SAP S/4HANA (OData) NetSuite (REST/SOAP) Odoo (XML-RPC / REST) Microsoft Dynamics 365 Oracle Fusion Cloud (REST) Custom REST & GraphQL SFTP feeds

What we need from you:

API endpoint URL

The base URL of your system's API (e.g., https://api.yourcompany.com/sap/odata/). Read-only access is enough.

API documentation

Link or PDF of your endpoint's docs so our developers know the available routes and response formats.

Auth details

OAuth2 / API key / Basic auth / mTLS — whatever your system uses. We coordinate credentials securely.

How to submit:

1Click Import data on the topbar.
2Choose Connect via API.
3Provide the API endpoint URL, documentation link, and auth type. Add a contact name + email so our developers can follow up.

What our developers do after receiving the request:

1Email you within 1 business day to coordinate sandbox credentials and confirm the data fields we'll pull.
2Build the connector on the Solomon side: authentication, field mapping to our IMS/WMS/TMS/FAMS schema, error handling, retries.
3Schedule automated ingestion (default: every 15 minutes) and set up monitoring + alerts.
4Run validation against your live data and walk you through the first day of automatic syncs.

Typical setup time: 3-5 business days from the day you share the endpoint + credentials. Once live, Solomon pulls data automatically — no more manual uploads on your side.

08 Troubleshooting Common fixes

Day counter looks wrong

Run Reset to empty before importing. The counter is based on distinct event dates in inventory data.

Findings disappeared

If you clicked Run all agents, that's a destructive refresh. Re-import the same day's CSV to bring them back.

"Rules" pill instead of "Solomon AI"

Configure your Solomon AI key in .env (variable GROQ_API_KEY), then run php artisan config:clear.

Imported but no changes show

Hard reload your browser (Ctrl+Shift+R). Old JavaScript may be cached.

"PO not found" skip on po_delivery rows

Expected on the first import; Solomon auto-creates the PO record. Subsequent imports for the same PO will update its status.

Solomon Agentic v0Pilot guide & demo handbook

Solomon Agentic v0
Pilot guide & demo handbook