AlttworkFinBook a demo
Alttwork Fin · conversational reports & dashboards
AI-native · self-serve · explainable

Describe the report. Skip the engineering ticket.

Alttwork Fin turns plain-language requests into validated reports and dashboards — correct data source, query, format and schedule — with chat, a customise panel, and a SQL editor all writing to one live preview. No implementation handoff.

From entry to scheduled delivery

The same refinement loop powers reports and dashboards — and nothing resets between turns.

01

Entry

Click “+ New”. A chat panel and a live preview open side by side.

02

Describe

Type in plain language or pick an AI-suggested metric for your role.

03

AI config

Query generated, output/chart suggested, preview renders on real data.

04

Refine

Chat, customise panel, or SQL — all sync to the same preview.

05

Save

Save, schedule, and deliver — with conditional triggers and audit.

Everything in the spec, in one agent

Ten reporting capabilities and seven dashboard capabilities, on a shared conversational core.

New report · chat
—
High-level architecture

How the platform works

A conversational experience on top of an AI orchestration layer, a validated config + semantic layer, and a governed execution & delivery engine — with security and audit running across all of it.

L1

Experience layer

Chat panelLive previewCustomise panelSQL editorSchedule & deliveryUpload detect
L2

AI orchestration / agent

Intent & parameter extractionClarification (one question at a time)Config synthesiserSQL generator + validatorChart-type recommenderUpload analyserExplainabilitySession state & sync
L3

Config & semantic layer

Validated config storeSchema catalog / data dictionaryVM templates & headersRole template library
L4

Execution & data layer

Query engineTenant data sources (ledger · settlement · recon)Render engine (table · charts · VM)Export engine (xlsx · csv · pdf · png)
L5

Delivery & automation

Scheduler (CRON · TZ)Conditional triggersEmail delivery (CC/BCC · retries)File naming & collision

Governance & security

  • Field masking & AES-256 encryption
  • Tokenisation vault per tenant
  • RBAC role-scoping; draft → publish
  • Every AI action explainable
  • Full audit: edits, exports, trigger evals
  • Accept / reject / modify any suggestion

Request lifecycle

1
Intent

NL prompt or uploaded sample

2
Extract + clarify

Params resolved; one question if unclear

3
Synthesise

Config + SQL generated & validated

4
Execute

Query runs; preview renders on real data

5
Refine

Chat · customise · SQL → one preview

6
Deliver

Schedule, triggers, export, audit