CommercialAutoRecon · reconciliation desk & MIS for financial institutions

Close the day before the day starts.

For finance, operations and treasury teams at banks, NBFCs and payment companies. Every recon — UPI, cards, NEFT/RTGS, nostro, vendors — matched daily on rules you write; breaks explained and aged; the MIS and board pack drafted with every number traced to the query that produced it.

  • 85–95% auto-match target
  • Every number traced
  • −50–70% pack prep target
TodayMISOps
AutoRecon: Every match run with records, match rate and breaks — and why the breaks happened.
Product screen with illustrative data.

WhySound familiar?

Three things we hear in every first conversation.

From finance, operations and treasury heads at banks, NBFCs and payment companies. Composite quotes from discovery calls — the problems are real; the names are left out.

AutoRecon: The problem, before AutoRecon
Three days before the board meeting
  1. “Month-end is a week of spreadsheets.”

    Recons are matched in macros that one person understands; breaks age until someone notices.

    Head of operationsMonth-end crunch

  2. “We do not know why the breaks happen, only that they do.”

    The same breaks recur every day because nobody turns them into a rule.

    Recon leadRecurring breaks

  3. “The board pack takes six days of copy-paste.”

    Numbers are pulled from five systems by hand, and every change means doing it again.

    CFOPack prep eats the week

HowWhat changes with AutoRecon

The same three problems, on the screens that solve them.

Product screens with illustrative data. In a working session we run one day of your files.

01Spreadsheet matching

Every recon matched daily, on rules you write.

  • Exact, tolerance and fuzzy narration matching
  • Breaks classified, aged and owned

Close every day, not every month

AutoRecon screen: Every recon matched daily, on rules you write.

02Recurring breaks

A recurring break becomes a suggested rule.

  • The pattern explained, the rule proposed
  • Maker-checker before it takes effect

Breaks that stop recurring

AutoRecon screen: A recurring break becomes a suggested rule.

03Copy-paste packs

The pack drafted, every number traced.

  • Each section linked to the query that produced it
  • Changes since the last pack highlighted

Pack prep in hours

AutoRecon screen: The pack drafted, every number traced.

FieldAutoRecon in the field

The people, the phone, the place.

The workplace it runs in, the person who uses it on a phone, the team at the desk and the detail it reads.

AutoRecon: Where AutoRecon runs
Where AutoRecon runs
AutoRecon: On the phone, where the work happens
On the phone, where the work happens
AutoRecon: The team at the AutoRecon desk
The team at the AutoRecon desk
AutoRecon: The detail AutoRecon reads
The detail AutoRecon reads

Photographs are illustrative — real customer sites are not shown.

  • 85–95%Auto-matchtarget on high-volume recons
  • 8–10 weeksTo first valueon one recon
  • DailyClosenot a month-end crunch
  • ExplainedBreakswith a suggested rule
  • −50–70%Pack prepMIS and board packs
  • Maker-checkerOn every ruleone audit log

01The screens

From a million transactions to a close the checker can sign.

The recon desk, the board pack and the morning close.

autowhat · AutoRecon / Reconciliation desk Illustrative data
AutoRecon screen

Every match run with records, match rate and breaks — and why the breaks happened.

Product screens with illustrative data — every name, place and figure is a placeholder.

02Before and after

Spreadsheet matching and a month-end crunch.

Recons are matched in spreadsheets, breaks age quietly, and the board pack takes a week of copy-paste.

AreaTodayWith AutoRecon
MatchingSpreadsheets and macrosRules you write, run daily
BreaksAge quietlyExplained, aged and owned
RulesIn one person’s headVersioned, with maker-checker
MISCopy-paste for a weekDrafted with every number traced
Regulatory returnsRebuilt each timeDrafted from the same record

03Match, explain, report

Eight steps, from raw files to a signed pack.

Read-only by default: AutoRecon reads your files and systems and writes nothing to your core unless you open the path.

  1. 01

    Ingest

    Files · APIs

    Switch, network, core, GL, bank and MT statements loaded as they land.

  2. 02

    Normalise

    Formats

    Every source mapped to one schema; bad rows rejected with a reason.

  3. 03

    Match

    Your rules

    Exact, tolerance and fuzzy narration matching.

  4. 04

    Explain

    Breaks

    Each break classified, aged and given an owner.

  5. 05

    Suggest

    Rules

    Recurring breaks become a suggested rule for the checker.

  6. 06

    Close

    Daily

    The close approved by maker and checker.

  7. 07

    Report

    MIS · board

    Packs drafted with every number traced to its query.

  8. 08

    File

    Regulatory

    Returns drafted from the same record, checker before submission.

04The platform

Recon, rules, reporting and controls.

Start with the noisiest recon; the rest join on the same engine.

Recon · 4

  • UPI & IMPS
  • Cards & network
  • NEFT / RTGS
  • Nostro & vendors

Rules · 4

  • Rule library
  • Fuzzy narration
  • Tolerances
  • Suggested rules

Reporting · 4

  • Daily close
  • MIS
  • Board pack
  • Regulatory returns

Controls · 4

  • Maker-checker
  • Break ageing
  • One audit log
  • Read-only default

05How we deliver

One recon, one metric, eight to ten weeks.

On the files and systems you already have.

  1. Weeks 1–2 · MapSources, formats and current rules mapped; match-rate baseline taken.
  2. Weeks 2–7 · BuildIngestion, matching rules, break workflow and the daily close.
  3. Weeks 7–10 · Go liveOne recon live in parallel, then primary.
  4. Then · ExtendOther recons, MIS and the board pack.

FAQQuestions we get about AutoRecon

Straight answers, before the call.

AutoRecon is an Autowhat AI product for finance, operations and treasury heads in financial services. Short, factual answers for buyers, procurement and the assistants people ask first.

In short. AutoRecon is an Autowhat AI product for finance, operations and treasury heads in financial services. It is built for finance, operations and treasury heads in financial services. It exists to remove the problems those teams describe in almost every first conversation: a week of spreadsheets at month-end; the same breaks recurring every day; and board packs that take six days of copy-paste. With AutoRecon, teams get UPI, cards, NEFT/RTGS, nostro and vendor recons matched daily, recurring breaks turned into suggested rules, maker-checker, and packs drafted with every number linked to its query. It connects to Switch & network files, Core & GL, MT950 statements and ERP, is used by Recon · finance · treasury, and runs on web + whatsapp. Like every Autowhat AI product, AutoRecon goes live on one site, lane or process first, usually within weeks, on managed cloud, in your own VPC or on-premises, and shares one operational record with the other products, so the second deployment reuses the data and integrations of the first.

Who is AutoRecon for?

Finance, operations and treasury heads in financial services. Daily reconciliation across switch, network, core, GL and bank statements on rules you write; breaks explained, aged and owned; MIS, board packs and regulatory returns drafted with every number traced to its source.

What problems does AutoRecon remove?

a week of spreadsheets at month-end; the same breaks recurring every day; board packs that take six days of copy-paste.

What do we get with AutoRecon?

uPI, cards, NEFT/RTGS, nostro and vendor recons matched daily; recurring breaks turned into suggested rules, maker-checker; packs drafted with every number linked to its query.

What does AutoRecon connect to, and who uses it?

It connects to Switch & network files, Core & GL, MT950 statements, ERP. Users: Recon · finance · treasury. Channel: Web + WhatsApp.

How do we start with AutoRecon?

With a 30-minute working session on your own data — an engineer from the relevant sector, not a sales queue. The first wave is scoped in that call; most products go live on one site, lane or process within weeks and then scale.

Where does AutoRecon run, and who owns the data?

On managed cloud, in your own VPC (AWS, Azure or GCP) or on-premises, with your keys. SOC 2 Type II and ISO 27001 controls, India data residency when required, and local or open-weight models when policy demands it. The data stays yours.

06Working session

Bring one day’s files. We’ll show you the match rate before lunch.

Thirty minutes: your noisiest recon, its sources and rules — and a first scope.