The flagship — everything else arranges behind it

Dovrin Toolbox. The dashboard you open every day.

A user dashboard for any M3 ERP product — with ticketing integrations and working knowledge of the M3 ION API suite, bringing Consequence Preview and AI. Daily use is where trust is built or destroyed: a wrong answer inside a tool someone opens forty times a day is a credibility problem that never goes away. That's why the Toolbox carries the story.

Watch the live demo → What's inside
Built for Infor M3 CloudSuite M3 Agentic Solutions — powered by Dovrin Narrate + Consequence Preview included, not separate SKUs
What's inside the Toolbox

Narrate and Consequence Preview aren't separate products. They're what the Toolbox does.

Whichever repo a capability lives in is an implementation detail. What matters to the person using it: ask a question in the dashboard you already have open, and get a grounded answer or a governed write — never a guess.

Capability

Narrate

Grounded, source-cited narration and briefs. Every number carries where it came from; thin or ambiguous data gets an honest refusal instead of a confident paragraph.

  • Two lenses. The same period narrates differently for a CFO and an AP manager — same facts, different framing.
  • Withheld, not guessed. If a record's status doesn't clear the confirmed-open bar, Dovrin says so instead of stating a balance it can't cite.
Capability

Consequence Preview

The Toolbox's core mechanic, and the single best live demo in the portfolio — watch it tell you what will happen before you let it happen.

  • Assembled from live reads. Every downstream impact is traced to the read that produced it — held orders, exposure, policy.
  • Approve an outcome, not a diff. The human sees consequences, not a raw field change, before confirming.

Ticketing integrations

The Toolbox is built to sit alongside the ticketing systems a team already runs, not replace them.

ION API working knowledge

Fluent in the M3 ION API surface so a request maps to the right program and transaction, not a guess.

Human-confirmed writes

An agent can propose a write, never execute one. A verified operator confirms; it runs exactly once.

Why it's trustworthy

The Toolbox is the newest place the discipline shows up.

The migration accelerator proved it first. Here's the same discipline, in the tool you'd actually live in.

Pillar 01

Every claim is sourced, or it's withheld. Never guessed.

Seven source tiers, enforced in code, from live-tenant confidence down to UNVERIFIED — a hard stop. The answer is withheld, not degraded.

Pillar 02

Every write is human-reviewed, with a record that can't be quietly edited.

A draft carries no token. A human approves at a dedicated confirm surface, which mints a one-time token server-side — the agent never holds it. The result lands in a tamper-evident, hash-chained ledger.

Pillar 03

When our own governance is broken, we say so — dated, before a customer finds it.

2026-06-10 An adversarial audit found the write gate's "triple lock" was, as wired by default, a double lock. Fixed and published before any tenant.
2026-08-30 A portfolio-wide verification pass found we'd undercounted our own test suite by ~13% and caught a CI failure nobody's build had surfaced. Fixed the same day.
The governed rail underneath — containerized create → rehearse → verify, tamper-evident hash-chained audit — is why this ledger is credible rather than a log file someone could edit afterward. It's built and gated (173 tests, image-gated) on a mock adapter; live-tenant wiring is the next milestone, cited here as plumbing, not pitched as a separate product.
See it work

Watch the mechanic that makes this real.

Consequence Preview and Narrate both run as live interactive demos on the homepage — a governed write from ask to audit, and a grounded narration that refuses before it guesses.

The interactive demo lives on the homepage

Ask Dovrin to raise a credit limit and watch the read → preview → confirm → audit loop run end to end, or ask it about overdue AP and watch it withhold rather than guess.

Open the demo →
Design-partner program

Put the Toolbox on top of your M3 — with a partner, not a pilot light.

We're onboarding a small number of design partners. You bring the M3 environment and the use case; we bring the governed engine and build it with you.

Request design-partner access Replay the demo
Honest status: the Toolbox is design-partner-ready, not yet generally available. The governed loop runs end-to-end on representative data; live-tenant validation is exactly what the design-partner program is for.