Ongoing live pilot · since

From unread query to review-ready response.

Mach Lilies AI Helpers monitor unread messages in 1-800Accountant’s Queries workflow, recognise account-validation and missing-information requests, and prepare draft responses for the client team.

1-800Accountant

  • 67% less Reported time spent Internal pre/post staff comparison
  • $666K Estimated annualised gross staff-capacity value Client labour-hour model
  • 1 inbox Existing Queries workflow Human-controlled drafts

Based on a client-reported internal pre/post comparison and client labour-cost assumptions. The $666K figure is annualised gross staff-capacity value, not an independently audited cash saving.

See the workflow Explore all results

01 · The client

A high-volume virtual accounting workflow.

1-800Accountant combines accounting technology with human oversight for small businesses, freelancers and startups. Its Queries workflow receives customer messages that can require account validation or requests for additional information before the team can continue.

02 · The operational problem

Opening the email was not the work.

  1. Unread query
  2. Understand what the customer needs
  3. Identify an account or information issue
  4. Collect the relevant context
  5. Draft the next response
  6. Return the case to the team

Repetitive handling time accumulated in the interpretation and response work around each query. The opportunity was not to automate professional judgement; it was to prepare the next safe, reviewable action sooner.

03 · Before and with

The same queue, run two ways.

Before: a person opens the unread query, determines the request type, checks what account or information context is missing, drafts the response, and the workflow continues manually. With Mach Lilies: the Helper monitors the agreed unread-query route, recognises the account-validation or missing-information context, and prepares a contextual draft; the client team then reviews and controls the next client-facing action.

Before

  1. A person opens the unread query.
  2. The person determines the request type.
  3. The person checks what account or information context is missing.
  4. The person drafts the response.
  5. The workflow continues manually.

With Mach Lilies

  1. The Helper monitors the agreed unread-query route.
  2. The Helper recognises the account-validation or missing-information context.
  3. The Helper prepares a contextual draft.
  4. The client team reviews and controls the next client-facing action. People decide

The Helpers prepare drafts; people retain control of client-facing decisions.

04 · Inside the workflow

Watch a query become a review-ready draft.

The reconstruction below follows one synthetic case through the stages the Helpers actually work: an unread message in the Queries inbox is recognised, the missing information is identified, a draft is prepared, and the case waits for a person.

Sanitised workflow reconstruction · illustrative interface · no customer data.

Queries inbox

  • Customer 024 Account information required Draft prepared
  • Customer 031 Validation check requested Awaiting reviewer
  • Customer 017 Further information needed Unread

Case rail · Customer 024

  1. Issue recognised Account-validation request identified
  2. Missing information identified Required account details not on file
  3. Draft prepared Contextual response ready for review
  4. Human review Awaiting reviewer

Draft response · Customer 024

Prepared by the Helper · awaiting reviewer · nothing sends without a person

05 · Reported results

Less handling time, expressed honestly.

Reported query-handling time, indexed

Reported baseline time index
100
Reported time index with the Helpers
33

Difference: 67% lower

Indexed visual derived from the client-reported 67% reduction. It does not represent absolute minutes or system-timestamp data.
$666K Estimated annualised gross staff-capacity value

The client’s internal labour-hour model valued the reported time release at approximately $666K on an annualised gross-capacity basis. This describes the estimated value of released staff time; it does not state that payroll or another cash expense fell by $666K.

Live since Ongoing proof-of-concept deployment
06 · Who does what

AI-supported. Human-owned.

AI Helper scope

  • Monitor the agreed unread-query route.
  • Recognise account-validation and missing-information requests.
  • Prepare draft responses.
  • Keep the next action reviewable.

Human-owned scope

  • Decide whether the draft should be used.
  • Make accounting, tax and professional judgements.
  • Handle sensitive, unclear or exceptional queries.
  • Control client-facing decisions.
07 · Methodology

How the result was measured and described.

Relationship
Ongoing live proof of concept
Deployment start
4 November 2025
Workflow
Queries inbox response drafting
Primary result source
Internal client pre/post staff comparison
Reported outcome
67% less time spent on the defined query-handling workflow
Financial classification
Estimated annualised gross staff-capacity value
Source currency
USD
Audit status
Not independently audited

The result is specific to this workflow, client environment and internal comparison. It is not a guarantee of the same result for another practice.

What this proves

Not an MTD case study. A proven operating pattern.

This deployment did not concern MTD quarterly record chasing. It demonstrates the underlying inbox pattern that Quiet Quarter applies to UK accountancy practices: identify the request, recognise missing information, prepare the next action and keep a person in control.

Read How Quiet Quarter is scoped for accountancy practices, or see both deployments side by side.

The next useful step

Bring us one chase loop.

We will tell you whether it is worth installing, what can move safely and what must stay human-led.

Book a fit call See all client results