AI that assists finance teams, with control over the final numbers.
AI agents sit inside each module and take the repetitive steps: matching, capture, prediction and first drafts. They propose; a named reviewer inspects the evidence and decides. Nothing posts on its own.
Ask in plain language
Questions answered from your own ledgers.
Every answer names its source report, so a reviewer can check it.
What the AI does
Three kinds of assistance, each with a clear owner.
Every capability is labelled by how it acts: automated processing, an AI suggestion, or a human-approved action.
AI-assisted processing
- Matching bank lines and intercompany balancesSuggestion
- Flagging unusual movements and reconciling differencesAutomated check
- Suggesting account mappings when data arrivesSuggestion
- Posting a match or journalHuman-approved
AI-assisted analysis
- Explaining period-on-period movementsSuggestion
- Identifying unusual balancesAutomated check
- Drafting management commentary from your dataSuggestion
- Publishing commentaryHuman-approved
AI-assisted documentation
- Preparing schedules and supporting explanationsSuggestion
- Drafting disclosure text where supportedSuggestion
- Listing missing information for reviewAutomated check
- Finalising a note or scheduleHuman-approved
One workflow, end to end
An intercompany mismatch, from detection to approval.
From the worked example: Apex Asia shows a payable of 0.38 to the parent, which shows a receivable of 0.40 (USD m).
- 1Automated check
The difference is detected
Intercompany matching finds a 0.02 difference between the parent and Apex Asia at period end.
- 2AI suggestion
The likely cause is explained
The agent finds a payment of SGD 26k from Apex Asia on 30 Sep that reached the parent on 1 Oct: cash in transit.
- 3Evidence
The supporting records are attached
Both ledger lines, the bank statement line and the proposed entry: Dr Cash in transit 0.02, Cr Intercompany receivables 0.02.
- 4Human-approved
An authorised reviewer decides
The reviewer inspects the evidence and approves, edits or rejects. Only then does the entry post, and the trail keeps both.
Agent library
14 agents, module by module.
Hover a module, or tap it on a phone, to see what its agents do. Every agent proposes; a named reviewer approves.
Close & reconcile
Match, flag and correct before the reviewer opens the period.
Transaction matching
Matches bank lines, intercompany entries and sub-ledgers to the GL, and explains what is left.
Anomaly detection
Flags unusual journals, balances and movements before review.
GL correction suggestions
Proposes the correcting journal for a reconciling difference, with the evidence it used.
Comply & AI
Keep every account tagged to the standard that applies.
Standards tagging
Suggests the IFRS 18 or ASU 2024-03 category for each account, for the reviewer to confirm.
Order to cash
Faster credit reviews, sharper collections, cash applied on arrival.
Financial statement extraction
Pulls figures from customer financials for credit reviews.
Payment date prediction
Predicts when each customer will pay, from their payment history, and orders the worklist.
Remittance capture
Reads remittance emails and attachments and applies cash to the right invoices.
Procure to pay
Invoices read, checked and coded before anyone opens them.
Invoice data capture
Reads invoices from the AP inbox and portal and extracts header and line data.
Duplicate and fraud screening
Checks each invoice against existing invoices and supplier details before approval.
Non-PO invoice coding
Suggests the GL account and cost centre for invoices without a purchase order.
Supplier statement reconciliation
Reconciles supplier statements to the ledger and lists the differences.
Treasury
A 13-week forecast that builds itself from the ledger.
Cash forecast builder
Builds the 13-week forecast from predicted receipts, scheduled payments and the plan.
Plan & analyse
First drafts of commentary, and answers with their sources.
Commentary drafting
Drafts flux and ratio explanations for the reviewer to edit and approve.
Finance assistant
Answers plain-language questions from your own data, with the source of every figure.
Governance
Propose. Review. Post.
The agent proposes
A match, a code, a correction or a draft, with the evidence it used.
A named reviewer decides
Approve, edit or reject. Thresholds decide when a second reviewer is needed.
The trail keeps both
What was proposed, what changed and who approved, kept with the period.
FAQ
AI questions
Anything else, ask us in the demo. You will be talking to the people who build it.
Does the AI post anything on its own?
No. Every proposal needs a reviewer's approval before it posts.
What data do the agents use?
Your own ledger, sub-ledgers, bank lines, invoices and remittances within FinConso, under the same access controls as your users.
Can we turn agents off?
Yes. Each agent can be enabled or disabled per entity and module.
Book a demo
See the consolidation workflow, from trial balance to signed statement.
A 30-minute walkthrough with the people who build FinConso.
- Entities, eliminations, translation, NCI and the multi-GAAP layer, run live on a sample group.
- Every consolidated figure drilled back to the entity and account it came from.
- Questions on IFRS, US GAAP, Ind AS or SFRS(I) answered by the team that builds the product.
- Want to test it on your own numbers? See how the structured pilot works.
Online booking opens soon.
The booking form is being set up. Please check back shortly.