FIS launches Agentic Insights Suite for cash forecasting and risk analysis

Sourced news briefing

FIS announced Agentic Insights Suite on 12 October 2026, describing a finance tool that combines internal and external data for cash forecasting and risk analysis. The release includes a new forecasting module and distinguishes planned payment anomaly and dispute forecasting features from the current launch. Company claims are not verified performance guarantees. The independent merchant analysis below focuses on how a business could evaluate forecasting assistance while retaining control over its payment and treasury decisions.

Independent merchant analysis: distinguish sales from available cash

A merchant can report strong sales while having limited money available to meet immediate obligations. Authorised payments, captured transactions, receivables, settlement balances and bank cash represent different stages. A forecasting project should define each stage before connecting systems. If the model treats every sale as immediately available cash, a sophisticated interface will produce a misleading answer. The underlying definitions matter more than the apparent precision of the forecast.

Consider a hypothetical subscription business with monthly customer charges and weekly supplier payments. Some customer payments may fail, some may arrive later than expected and some may be refunded. The business needs a forecast that reflects those differences rather than simply dividing monthly revenue across calendar days. The example illustrates a modelling requirement, not a claim about the capabilities of any particular product or a recommendation to change billing terms.

Settlement deductions also need consistent treatment. Finance should distinguish fees, reserves, refunds, disputes and adjustments rather than placing every difference into a generic variance category. Each has a different implication for timing and availability. The forecast should trace back to the merchant's actual reports and agreements. A model cannot resolve an undocumented settlement assumption merely by incorporating more external information.

Independent merchant analysis: make the data explainable

Start with a source inventory. Identify which system owns invoice status, payment events, customer balances and bank receipts. When two systems disagree, decide which record controls and how the discrepancy is investigated. Otherwise automated analysis may combine contradictory information without anyone understanding the resulting number. The project should improve the reliability of finance records rather than building a more elaborate layer above unresolved inconsistencies.

Data freshness should be visible to the person using the forecast. A bank feed from this morning and a receivables extract from last week do not describe the same moment. The interface should make those differences understandable. If an important source fails to update, users need a clear warning and a defined fallback. Silent continuation can make an outdated forecast look current simply because the dashboard itself refreshed.

Preserve explanations for unusual movements. A large settlement might reflect delayed receipts from a prior period rather than a surge in sales. A sudden outflow may be a planned supplier payment rather than a deterioration in trading. Staff should be able to attach the relevant business context and later assess whether it improved the forecast. Human knowledge is useful when it is recorded consistently instead of remaining in private spreadsheets or conversations.

Independent merchant analysis: evaluate uncertainty honestly

Forecasts should support decisions under uncertainty rather than disguise uncertainty with a single number. Ask how the system represents delayed receipts, changing refund levels or a temporary payment interruption. Compare a base case with plausible adverse conditions relevant to the business. The purpose is to identify which commitments become difficult under those conditions, not to present a simulation as a prediction that those events will occur.

Evaluate accuracy at the horizon that matters to the decision. A model might estimate monthly receipts reasonably while performing poorly on the specific day payroll is due. Conversely, a short horizon forecast may be useful for routine liquidity but insufficient for a planned expansion. Define the decision first, then choose the measurement period. Report errors in a way that shows their operational consequence rather than only a statistical average.

Backtesting should use information that would actually have been available at the time. If later settlement data enters an earlier forecast, apparent performance can become misleading. Keep a dated forecast record and compare it with subsequent outcomes. Investigate material misses by cause, such as incomplete data, timing assumptions or genuinely unexpected events. That review helps determine whether the model, the data process or the business assumption needs correction.

Independent merchant analysis: separate advice from execution

A recommendation to accelerate collections or defer a payment can affect customers and suppliers. The business should define which actions require approval and which may be performed automatically. A forecasting tool does not need unrestricted authority over bank accounts to provide useful analysis. Keep permissions proportionate to the task and ensure that staff can understand the consequences before authorising a material action.

Collections decisions need commercial context. An overdue invoice may be disputed, awaiting a contractual milestone or already covered by a negotiated arrangement. Automated prioritisation should not cause staff to contact a customer with an incorrect demand. Connect relevant case information and preserve the reason for exceptions. The objective is more informed action, not simply a larger number of collection messages generated from an incomplete receivables list.

Supplier relationships require similar care. A model may identify an apparent cash benefit from delaying an outflow, but the payment could support essential inventory or a time sensitive service. The person deciding should see the operational context alongside the financial estimate. This does not invalidate forecasting; it defines its role as one input to a controlled business decision rather than a substitute for accountable management.

Independent merchant analysis: govern the implementation

A pilot should have a clear boundary, such as forecasting one entity's settlement receipts without moving funds. Compare outputs with the existing process and record both improvements and unresolved differences. Assign owners for data quality, forecast review and access permissions. A shared project without named responsibility can leave finance assuming technology owns the numbers while technology assumes finance has already validated them.

Vendor diligence should include retention, access, export and continuity questions for the data being supplied. Finance records can reveal customer relationships and business performance. Ask who can access them and how the business would retrieve useful records if it changes providers. These are implementation questions to resolve through the proposed service and contract, not claims that the announcement discloses every relevant protection or operating detail.

Keep the evaluation grounded in work saved and decisions improved. Measure time spent reconciling sources, explaining forecast changes and preparing liquidity reviews. A compelling demonstration may answer natural language questions quickly while leaving the underlying preparation unchanged. The business case should show whether the new process reduces avoidable effort and produces information that management can use with an appropriate understanding of its limitations.

Create a practical handover for absences. Another authorised finance colleague should be able to locate the latest forecast, understand material assumptions and identify outstanding data issues. An analytical tool should reduce dependence on undocumented personal knowledge. That benefit is easier to sustain when review notes and decisions remain available through an agreed shared process.