Intelligence that executes, learns, and anticipates across your entire treasury
Meet TellMe, Embat's AI agent built for corporate treasury. It works across every process: treasury, accounting, payments, and cash optimisation. It handles the operational work your team shouldn't do manually and surfaces the decisions that need your judgment.
How TellMe works
Autonomous when it can be. Guided when it should be.
Silent Mode
Routine tasks run in the background: categorisation, enrichment, reconciliation. Executed, logged, reversible. Nothing happens without you being able to see it and undo it.
Guided Mode
When TellMe detects something that needs your judgment, it presents it with context and reasoning. No fluff, no guesswork. Approve, modify, or reject in one click.
Ask Mode
Request anything: a cash flow breakdown by entity, an explanation of why last week's forecast missed, a bulk recategorisation. You ask, TellMe delivers. Same transparency, same control.
The intelligence behind Embat
What does TellMe do?
Everything your finance department needs:
Smart categorisation
Categorise every transaction automatically, from day one
Every incoming transaction is categorised automatically, with accuracy improving as your patterns are learned over time. Your financial picture stays current so your team makes decisions based on real data, not guesswork.
Invoice enrichment
Automatically turn raw forecast imports into structured, categorised data
Categories are assigned and each record linked to its corresponding bank product before you need it. No manual clean-up. No reconciliation surprises.
Forecast reconciliation
Match every forecast to its transaction with context-aware reconciliation
Every transaction is cross-matched against its corresponding forecast entry, generating a clear picture of what was expected, what arrived, and what is still pending. Every correction you make improves its logic continuously.
Payment behaviour analysis
Adjust forecast dates to match how your counterparties actually pay
Historical payment patterns across all your counterparties are analysed, flagging forecasts whose dates don't reflect real behaviour. Accept the proposal and your cash flow projections will immediately get sharper.
Forecast gap detection
Fill the gaps in your cash flow with forecasts based on historical patterns
Gaps in your forecast coverage are identified continuously. When one is found, a forecast grounded in your historical data is suggested, keeping your cash flow view complete and ready for decision-making.
Cash flow projection
Project cash flows automatically, even where historical data is missing
Categories lacking a direct data source are automatically identified, with forecasts generated from historical trends. You can apply manual adjustments at any time to ensure future estimates remain accurate.
Reporting intelligence
Turn your treasury reports into instant insights, complete with narrative
Your reports are automatically interpreted, with key insights highlighted and professional commentary added to provide context. Reporting cycles that once took hours are now completed in minutes.
Invoice matching
Settle invoices automatically through bank transaction matching
Each incoming transaction is analysed and matched to the right invoice automatically, whether it's a 1-to-1 match, multiple invoices, or a partial payment. Accept in one click and move on.
GL account posting
Post transactions directly to the right GL account, automatically
When there are no associated invoices, the description, counterparty, and metadata are analysed to suggest the most accurate GL account, cost centre, and analytical tags. Every transaction gets a proposal grounded in your own accounting logic.
Bank reconciliation
Reconcile your bank statements and GL entries in real time
Every transaction posted through Embat is automatically reconciled with its corresponding journal entry. For entries posted outside Embat, matching transactions are identified based on metadata and surfaced as suggestions.
PSP order matching
Automatically match payment gateway transactions to your sales orders
Patterns between each PSP transaction and its corresponding sale are detected automatically, performing matchings based on ID, amount, and date. Your accounting team only sees what actually needs attention.
PSP order matching
Resolve PSP discrepancies with intelligent ledger account postings
When a transaction and a sale share the same ID but amounts don't align, the difference is analysed and a posting against the appropriate ledger account is suggested. Every discrepancy gets a clear, actionable proposal.
Liquidity optimisation
Optimise your liquidity allocation across all entities and currencies
Your cash position across all entities and currencies is analysed automatically, with optimal allocation moves suggested to maximise returns and minimise costs. All transactions are reconciled through the intercompany and accounting module, closing the loop without manual intervention.
Payment data completion
Complete payments instantly: TellMe fills in the missing beneficiary data
As you build your payment batches, incomplete records are detected in real time and the missing beneficiary data suggested on the spot. Keep your payment workflow moving without interruptions or manual lookups.
Ask mode, everywhere you work
Work on your treasury from inside Claude
TellMe's Ask mode now connects to Claude via MCP. Authenticate once with your existing Embat credentials, ask in plain language, and get cited answers pulled from your live treasury. No dashboards, no exports, no tab-switching. Permissions stay scoped exactly to what each user can already see.
Live, cited data
Every answer traces back to the originating bank account or forecast entry, reflecting your last sync (not end-of-day batches).
Permissions enforced at the data level
Even if the model writes the query, it can't return rows the user isn't entitled to see. ISO 27001, SOC 2 Type II, GDPR.
Human always in control
Runs exclusively on Ask mode: every interaction is explicitly initiated, transparent, and reversible.
One platform. Every bank. Full control.
Actual results from real financial teams
This is how teams already using Embat describe the change.
"Previously, 100% of one team member's time was spent on manual tasks. Today, one person can manage daily treasury in real time, spending only 10–15% of their time on it, freeing up resource for value-added work."Javier Gaya - CFO at HOFF
"If you apply a dose of AI to this, the result is highly promising and exciting. Embat's Treasury and Accounting solution is perfect for our organisation."Ángel López, CFO, Grupo CPS
"We manage a wide range of B2C, B2B, and B2G transactions with customers in 60 countries. With Embat, we have been able to bring the month-end close forward by more than 4 days and we have centralised payments, reducing the time per transaction from 10 minutes to a matter of seconds."Ignacio Marcos Pérez, CFO, thePower
"A Treasury Management System makes sense if it enables the Treasury Manager and Corporate Finance to have clear visibility across three levels: the actuals, confirmed forecasts, and an estimated forecast for strategic decisions. For us, Embat has been the solution that ensures clarity across those three levels."Michele Laria, Corporate Finance Director, Molins
See TellMe in action
Ready to see what intelligence can do for your treasury?
In a focused demo session, see how TellMe executes reconciliation, forecasting, categorisation, and payment workflows across your entire treasury. From operational execution to strategic decisions, all under your control
FAQs about AI in Treasury and Finance
TellMe, Embat's AI agent, works across four treasury processes: treasury, accounting, payments, and cash optimisation. It categorises transactions from day one, reconciles bank statements against your ledger, adjusts forecast dates based on real counterparty payment behaviour, and detects gaps in your cash flow coverage. Every action is logged, reversible, and transparent.
TellMe operates in three modes: Silent (autonomous execution), Guided (suggestions that require your approval), and Ask (on-demand queries). You control how much autonomy it has at every step.
TellMe automates the most time-consuming accounting workflows: matching bank transactions against open invoices (including 1-to-many and partial payments), posting to the correct GL account based on learned patterns, reconciling bank statements against journal entries in real time, and matching payment gateway transactions to sales orders.
In treasury, it handles transaction categorisation, forecast enrichment, and cash flow projections. Your team shifts from processing every transaction to reviewing exceptions only.
Most finance teams see impact from day one. Silent Mode begins categorising transactions and reconciling accounts immediately, with accuracy improving as TellMe learns your patterns over the first weeks. Forecast adjustments based on counterparty payment behaviour are active within the first month.
TellMe operates within Embat's enterprise-grade security infrastructure: ISO 27001 certified, SOC 2 Type II audited, and GDPR compliant. All data is encrypted with AES-256 at rest and TLS 1.2+ in transit.
Customer data is isolated through database sharding, never mixed between organisations, and never used to train models without explicit consent. AI suggestions pass through business rule validation before reaching your team. Every action TellMe takes is logged, traceable, and reversible.