← Work index
CASE 03 / Operational AIPROTOTYPE

Avorth

From a useful answer to an accountable handoff

Two related builds: a read-only assistant over operational data, and a voice-intake workflow with durable records, exact-payload approval, and explicit handoff states.

CONTEXT
Startup / software built in 2026
TOOLS / METHODS
TypeScript · SQLite · Jobber · QuickBooks · Voice APIs · State machines
THE CENTRAL LESSON

A chatbot can answer “which jobs are uninvoiced?” Maintaining that condition over time also requires durable state, coverage checks, authorization, recovery, and a way to verify what happened outside the conversation.

2 evidence recordsRead the scope & limits ↓
THE QUESTION

Can an AI system maintain responsibility for an outcome instead of only answering a prompt?

Context

Avorth is one startup with two strands of work. The first connected a texting assistant to Jobber and QuickBooks data. The second connected voice-model intake to a local operator desk. Both expose the same design problem: the model’s output must pass through software that knows what it is allowed to do.

What I did

  1. 01

    Built an accounting-data adapter that rejects non-GET requests in code, even though the underlying accounting permission is broader.

  2. 02

    Tracked whether paginated reads reached the end, so an incomplete fetch cannot silently become a claim of complete coverage.

  3. 03

    Built a voice workflow that stores call evidence, prepares an exact action payload, and checks that payload’s hash at approval and execution.

  4. 04

    Separated attempted delivery from a named person’s acknowledgement. In the prototype, silence can exhaust the escalation chain without being labeled completion.

Architecture / workflow

AVORTH / SYSTEM STUDYINSPECTED PROTOTYPE SOURCE

Approval is not delivery. Delivery is not acknowledgement.

What would actually count as a completed handoff?

  1. 01
    MODEL ASSISTANCE

    Collect intake

    Voice-model intake produces a durable call record

  2. 02
    SOFTWARE LOGIC

    Bind the payload

    Prepare stored content and check its hash

  3. 03
    HUMAN JUDGMENT

    Approve that version

    Reviewer approves the exact payload

  4. 04
    SOFTWARE LOGIC

    Attempt the handoff

    Local desk paging and alerts are simulated

    SIMULATED / MODELED
  5. 05
    SOFTWARE LOGIC

    Check acknowledgement

    Track acknowledgement by a named person

INSPECT THE FAILURE BOUNDARYThe payload changes, or nobody acknowledges the attempted handoff.
WHAT FOLLOWS

Stale or corrupted payloads are rejected. Silence can exhaust escalation without being labeled completion. These are inspected prototype behaviors; the suite was not rerun for this site.

This drawing covers the voice prototype. The separate accounting adapter is read-only; it does not create invoices. No live dispatch or autonomous business outcome is implied.

Where it gets difficult

An approval becomes ambiguous if the content changes afterward. The prototype binds the reviewer’s approval to the exact stored payload and rejects a stale or corrupted one.

An API can return a valid page while leaving more records unread. Coverage state is part of the result, rather than a hidden implementation detail.

A delivery attempt, a delivery receipt, and acknowledgement are different events. The state model keeps that distinction visible to the operator.

THE BOUNDARY OF THE CLAIM

Limits & evidence

  • The local voice desk simulates paging and alerts. It is not evidence of a production dispatch or customer-notification system.
  • Provider demos and local implementations do not establish paying adoption or sustained business outcomes.
  • “Keep completed jobs from remaining uninvoiced” is the persistent-responsibility direction. The read-only adapter does not create invoices or establish that this condition was maintained autonomously in production.
Source inspection

Read-only adapters and coverage state

Inspected accounting and field-service adapters reject accounting writes and expose whether paginated reads reached the end.

What this does not establish

Private code and account records are withheld. Code inspection does not establish production adoption, complete coverage of every API, or a fresh test run.

Source inspection

Approval, delivery, and acknowledgement

Engine code and test assertions were inspected for exact-payload approval, stale or corrupted hashes, and acknowledgement versus exhausted escalation.

What this does not establish

Paging is simulated. Assertions were read, not rerun for this site. This is a prototype account, not an independent security audit or proof of live dispatch.

What this changed

A chatbot can answer “which jobs are uninvoiced?” Maintaining that condition over time also requires durable state, coverage checks, authorization, recovery, and a way to verify what happened outside the conversation.

WHERE THIS LEADS

How should a system preserve unresolved responsibilities across restarts and changes in people, tools, or data?

NEXT CASEVALUN