FIS Launches Agentic AI Cash Flow Forecasting: Finance Might Finally Get Ahead of the Money

FIS has introduced an AI-powered cash forecasting platform that promises to help CFOs see liquidity problems before they arrive. I think this could be one of the more useful applications of artificial intelligence in finance—not because it produces prettier forecasts, but because it could give companies time to make better decisions.

There’s a special kind of entertainment in watching a company announce record revenue while its finance team quietly wonders whether there’s enough cash to make payroll.

The income statement looks wonderful.

Sales are growing. Margins are improving. The CEO is discussing expansion.

Meanwhile, accounts receivable is aging like a particularly expensive cheese collection, three major customers haven’t paid, and someone has just approved a capital expenditure because the budget says there’s money available.

Technically, there is.

It’s just currently residing in somebody else’s bank account.

This is why I pay attention when technology companies announce improvements in cash flow forecasting and receivables collections.

Not because I believe every AI product is going to transform finance.

I’ve survived enough software demonstrations to know that the phrase seamless integration should occasionally come with a warning label.

But because cash forecasting is one of those areas where better information can lead directly to better financial decisions.

And FIS has just introduced something worth examining.

On October 12, 2026, financial technology company FIS announced its new Agentic Insights Suite, an AI-powered platform designed to help CFOs and finance teams manage collections, liquidity, credit risk and cash flow.

According to the company’s announcement, the suite includes cash flow forecasting across multiple time horizons, simulations of different cash outcomes, and the ability to combine internal financial information with external credit and public data.

It also supports natural-language questions and Power BI dashboards.

That’s an interesting combination.

Because the biggest problem with cash forecasting has never been our inability to produce another spreadsheet.

Finance departments are exceptionally good at producing spreadsheets.

We could probably forecast the weather using an Excel workbook if somebody gave us enough tabs.

The real problem is knowing what will happen to cash early enough to influence the outcome.

And that is where I think agentic AI could make a meaningful difference.

What Is FIS Agentic Insights Suite?

FIS Agentic Insights Suite is a newly announced platform for the office of the CFO.

It is intended to bring together financial and risk information that often lives in disconnected systems.

Think ERP data, treasury information, receivables activity, customer records and external credit indicators.

FIS says its new cash flow forecasting module can project cash across configurable time horizons and simulate different outcomes.

For example, a finance team could examine how delayed customer payments or faster-than-expected cash outflows affect liquidity.

The company also describes capabilities for assessing credit risk and business health using a combination of internal and external data.

The platform is designed to work with existing FIS receivables and treasury products, including GETPAID, BillerIQ, Integrity and Quantum.

And FIS says deployment can take weeks rather than months.

I find the direction interesting, although deployment speed and forecasting effectiveness will need to be demonstrated in actual customer environments.

The announcement describes what the platform is designed to do.

It doesn’t establish that every implementation will achieve a particular improvement in forecast accuracy or working capital.

That distinction matters.

Still, the underlying opportunity is substantial.

For years, many finance organizations have approached cash forecasting as a reporting exercise.

Collect the data.

Update the spreadsheet.

Compare the forecast with the actual bank balance.

Explain the difference.

Repeat.

The result may be accurate enough to satisfy a reporting requirement.

But I’d argue that the more valuable question is what finance can do before the cash moves.

The Real Value of AI Cash Flow Forecasting Is Earlier Decisions

Let’s imagine a company expects to receive $12 million from customers over the next 30 days.

The forecast includes several large invoices scheduled for payment.

The finance team uses historical payment patterns and customer terms to estimate collections.

Nothing unusual so far.

Now suppose one of those customers begins showing signs of financial stress.

Perhaps its payment behavior changes.

Perhaps external credit information deteriorates.

Perhaps the customer disputes a large invoice.

The traditional forecast may not reflect that development until someone manually updates the assumptions.

An AI-enabled process could potentially identify the change earlier and show how it affects expected cash receipts.

That creates time for action.

The collections team might contact the customer.

Treasury might adjust its short-term funding plan.

Management might delay a discretionary expenditure.

The company might arrange temporary liquidity before it becomes urgent.

None of those decisions requires a revolutionary new theory of finance.

They require information arriving sooner.

And that’s what makes this application interesting.

The financial value of AI cash flow forecasting may come less from predicting the exact bank balance and more from giving management additional time to respond.

A forecast that identifies a cash shortage three weeks in advance is more useful than a perfectly accurate report explaining the shortage after it happened.

I realize that sounds obvious.

But we’ve built a remarkable number of financial processes around explaining things that are already over.

How Agentic AI Could Improve the 13-Week Cash Flow Forecast

The 13-week cash flow forecast remains one of the most practical tools for managing short-term liquidity.

I like it because it forces finance to focus on actual cash movements.

Not revenue recognition.

Not EBITDA.

Not the amount management hopes customers will pay.

Cash.

A typical 13-week forecast includes expected customer receipts, payroll, supplier payments, taxes, debt service, capital expenditures and other cash movements.

The challenge is keeping the assumptions current.

A company may have hundreds or thousands of invoices.

Payment timing can change.

Purchase orders become invoices.

Invoices become payments.

Customer disputes delay collections.

And expenses arrive earlier than expected.

Updating all of that manually can consume a surprising amount of time.

Agentic AI could help by continuously incorporating new information and identifying changes that matter.

For example, an AI cash forecasting system might detect that several customers have begun paying later than their historical average.

It could then estimate the effect on weekly collections.

Instead of waiting until the next forecast cycle, finance could review the change immediately.

That’s a meaningful improvement.

Not because the machine has become the CFO.

But because the CFO doesn’t need to wait until Friday afternoon to discover that Wednesday’s assumptions are no longer true.

A Practical Cash Flow Forecasting Example

Let’s build a simplified example.

Assume a company begins the month with $5 million in cash.

It expects $8 million in customer collections and $10 million in cash outflows.

The original forecast looks like this:

Cash flow measure	Original forecast
Beginning cash	$5,000,000
Customer collections	$8,000,000
Operating and other cash outflows	($10,000,000)
Ending cash	$3,000,000

The company has a minimum operating cash requirement of $2 million.

That leaves $1 million of forecast headroom.

Management is comfortable.

Then a major customer delays a $1.5 million payment.

The revised forecast becomes:

Cash flow measure	Revised forecast
Beginning cash	$5,000,000
Customer collections	$6,500,000
Operating and other cash outflows	($10,000,000)
Ending cash	$1,500,000

Now the company is $500,000 below its minimum cash requirement.

The business hasn’t necessarily become less profitable.

The revenue may still be recognized.

The customer may still pay.

But the timing has changed.

And cash timing matters.

If finance identifies the delay early, management has options.

It could accelerate collections from other customers.

Delay nonessential spending.

Negotiate supplier payment timing.

Or arrange short-term financing.

If finance identifies the problem after payroll and supplier payments have already cleared, those options become considerably less attractive.

That’s the benefit I would want from AI-powered liquidity forecasting.

Not simply a new number.

A useful warning while management still has choices.

Working Capital Optimization Is a Better AI Use Case Than Another Dashboard

FIS cited findings from Deloitte’s Finance Trends 2026 survey showing that 46% of global finance leaders identified working capital optimization as a leading opportunity for agentic AI.

Yet only 14% reported having fully integrated AI agents into their finance function.

That’s a significant gap between interest and adoption.

And I think it makes sense.

Working capital is an attractive area for AI because improvements can often be connected to measurable financial outcomes.

Reduce overdue receivables.

Improve collection timing.

Identify payment disputes earlier.

Manage supplier payment schedules.

Reduce unnecessary cash buffers.

These activities can improve liquidity without requiring the company to sell more products.

Which is refreshing.

Not every financial improvement needs to begin with a revenue growth assumption that requires three new sales territories and an optimistic hiring plan.

Sometimes the money is already there.

The company simply needs to collect it.

How Much Cash Can a Reduction in DSO Release?

Let’s take a company with $120 million in annual credit sales.

Assume sales are evenly distributed throughout the year.

Average daily credit sales are approximately:

$120 million ÷ 365 = $328,767.

Now suppose the company reduces days sales outstanding, or DSO, by five days.

The approximate reduction in accounts receivable would be:

$328,767 × 5 = $1.64 million.

That’s approximately $1.64 million in cash released from receivables, assuming the sales level and other relevant conditions remain unchanged.

The company hasn’t increased its prices.

It hasn’t reduced headcount.

It hasn’t launched a new product.

It has improved the timing of cash collection.

Now consider what that cash could do.

Reduce borrowing.

Fund inventory.

Support capital investment.

Provide additional liquidity.

Or reduce the need for an expensive short-term credit facility.

If the company otherwise would have borrowed $1.64 million at an 8% annual rate, avoiding that borrowing for a full year could save approximately $131,500 in interest.

That’s a real financial benefit.

And it’s measurable.

This is why I think working capital optimization deserves serious attention in discussions about AI return on investment.

The benefits don’t have to be theoretical.

They can show up in the cash position.

Why Customer Credit Risk Belongs in the Cash Forecast

One capability in FIS’s announcement caught my attention: the combination of internal company data with external credit and public information.

I think that’s important.

A cash forecast often assumes that customers will pay according to expected patterns.

But customers aren’t all equally reliable.

A business with strong payment history may begin experiencing financial difficulties.

A customer may have increased its borrowing.

An industry may be facing weaker demand.

Or a major account may begin extending payment timing.

Those developments can affect cash collections before they become visible in reported revenue.

I’d want the finance team to understand that risk.

Suppose two customers each owe $500,000.

Customer A has a long history of paying within 30 days.

Customer B has recently begun paying after 60 days and has experienced deteriorating credit conditions.

A basic forecast might treat both invoices according to their contractual due dates.

A risk-adjusted forecast could treat them differently.

That doesn’t mean automatically assuming Customer B won’t pay.

It means using available evidence to develop a more realistic range of collection dates.

For a company with significant customer concentration, that information can be particularly valuable.

A single delayed payment may materially affect liquidity.

And I’d rather learn about the risk before the invoice becomes 90 days overdue.

Agentic AI Could Connect Treasury, Collections and FP&A

One reason cash forecasting becomes difficult is that the relevant information is distributed across several teams.

Accounts receivable understands customer payment behavior.

Treasury manages bank balances, financing and liquidity.

FP&A develops operating forecasts.

Accounts payable manages supplier obligations.

Sales may know when a major customer is experiencing difficulties.

Procurement may know when a significant payment commitment is coming.

Each team has part of the picture.

But the company needs one coherent view of cash.

That’s harder than it sounds.

Especially when systems don’t communicate well.

I think one of the strongest opportunities for AI is connecting these activities.

Imagine a collections team identifies a large payment delay.

That information updates the expected cash receipts.

Treasury sees the resulting liquidity impact.

FP&A incorporates the change into its rolling forecast.

Management receives an explanation of the financial implications.

The company can respond without waiting for several teams to reconcile different versions of the same information.

That would be useful.

And considerably better than discovering that treasury and FP&A have different cash forecasts because one team updated a file and forgot to tell the other.

I’ve always believed the phrase single source of truth should be treated as an aspiration until someone proves it.

What Makes Agentic AI Different From Traditional Cash Forecasting Software?

Traditional forecasting software can automate data collection, calculations and reporting.

Agentic AI generally refers to systems designed to perform more complex sequences of tasks, respond to changing information and potentially recommend or initiate actions within defined boundaries.

In cash forecasting, that could involve detecting changes in payment behavior, analyzing their impact and recommending responses.

The distinction isn’t simply that the software uses a machine learning model.

It’s how the system connects information to a workflow.

For example, a traditional forecast might show that expected collections declined.

A more advanced AI-enabled process could identify which customers contributed to the change, examine the likely reasons and highlight potential actions.

That’s the direction I find promising.

But I would distinguish between a system that recommends action and one authorized to execute it.

Those are different levels of responsibility.

For most organizations, I’d begin with analysis and recommendations.

Let the technology identify exceptions, explain changes and prioritize decisions.

Then establish appropriate approval processes before allowing it to initiate financial actions.

I’m enthusiastic about better forecasting.

I’m somewhat less enthusiastic about discovering that an automated agent decided to rearrange the company’s cash accounts while everyone was at lunch.

Controls still matter.

How I’d Evaluate the ROI of AI Cash Flow Forecasting Software

If I were evaluating a platform like FIS Agentic Insights Suite, I’d build the investment case around measurable improvements.

Not the number of AI features.

Not how impressive the dashboard looks.

And definitely not how many times the sales presentation uses the word transformation.

I’d begin with the financial outcomes.

How much time does the team currently spend preparing cash forecasts?

How frequently are forecasts updated?

What is the historical forecast error?

How much cash does the company maintain as a precaution against uncertainty?

What is the cost of short-term borrowing?

How much working capital could realistically be released?

And what would improved collections performance be worth?

Let’s consider a hypothetical implementation.

Assume a company expects the following annual benefits:

Potential benefit	Annual value
Reduced manual forecasting work	$120,000
Lower financing costs from improved collections	$130,000
Reduced external reporting and reconciliation costs	$50,000
Total annual benefits	$300,000

Suppose the software, integration and ongoing support cost $180,000 annually on an equivalent recurring basis.

That produces an estimated annual net benefit of $120,000.

A simple benefit-to-cost ratio would be approximately 1.67.

But I’d want to examine implementation timing, one-time costs and the likelihood of realizing each benefit.

I’d also distinguish cash savings from productivity capacity.

If employees save time but remain on payroll, that isn’t automatically a reduction in expenses.

It may create capacity for additional analysis or avoid future hiring.

Both can be valuable.

But I’d model them separately.

The objective is to establish a credible return.

A company shouldn’t need to invent spectacular benefits to justify useful technology.

Why Cash Forecast Accuracy Isn’t the Only Metric I’d Track

Forecast accuracy matters.

But I wouldn’t evaluate an AI cash forecasting platform solely on how closely its projections match actual bank balances.

I’d also examine whether the forecast helped management make better decisions.

For example, did the company identify cash shortages earlier?

Did it reduce emergency borrowing?

Did collections teams resolve overdue invoices faster?

Did treasury improve the timing of investment or financing decisions?

Did management reduce unnecessary cash reserves?

Those outcomes may matter more than a small improvement in statistical forecast accuracy.

Imagine two systems.

System A predicts month-end cash within 2% but doesn’t provide useful information until late in the month.

System B predicts within 4% but identifies a major collection delay three weeks earlier.

Depending on the company’s needs, System B may create more value.

That’s an important distinction.

The best cash forecast isn’t necessarily the one with the smallest error. It’s the one that helps the company make the best financial decisions.

I’d want both accuracy and actionable information.

But if I had to prioritize, I’d start with decisions.

The Data Problem Finance Still Has to Solve

There’s one issue that no AI product can completely eliminate.

The quality of the underlying financial information.

If customer records are inconsistent, invoices are incorrectly categorized or payment dates aren’t updated, the forecasting process will struggle.

An AI system may detect anomalies.

It may help reconcile information.

But the company still needs reliable data definitions, ownership and controls.

For example, what counts as an expected customer receipt?

An issued invoice?

An approved invoice?

A contractual payment date?

A probability-adjusted collection estimate?

Those definitions affect the forecast.

And if different teams use different assumptions, the system needs a clear method for reconciling them.

I’d want to understand that before implementation.

Who owns the forecast assumptions?

Who approves changes?

How are exceptions reviewed?

Can users trace a projected cash movement back to the underlying transaction?

And how does the company verify the model’s recommendations?

These questions aren’t reasons to avoid AI.

They’re the work required to use it effectively.

The technology may make the process faster.

Finance still has to make it trustworthy.

What I’d Want to See in an FIS Agentic Insights Suite Demonstration

If I were evaluating this product, I’d ask for a demonstration using realistic financial data.

Not a perfectly organized sample company where every customer pays on time and every system integrates beautifully.

I’d want something messier.

A customer with disputed invoices.

A supplier payment scheduled earlier than expected.

A large collection that slips into the following month.

Several currencies.

And a cash forecast that needs to be updated without rebuilding the entire model.

Then I’d ask the system to explain what changed.

Which assumptions moved?

Which transactions caused the change?

How much did the expected minimum cash balance decline?

What alternative outcomes should treasury consider?

Can the team trace the answer back to source data?

And can a human override an assumption with a documented explanation?

I’d also want to know how the system handles missing information.

Does it flag uncertainty?

Does it use historical patterns?

Does it ask for clarification?

Or does it produce a confident answer because confidence looks attractive in a software demonstration?

I’d prefer a system that admits uncertainty.

Finance has enough people producing confident forecasts already.

Why This Could Be Good News for Smaller Finance Teams

One aspect of AI cash forecasting that interests me is its potential usefulness for finance teams without enormous resources.

Large corporations may have dedicated treasury departments, sophisticated forecasting platforms and specialized analysts.

Smaller companies often rely on a handful of people managing several responsibilities.

The controller may oversee cash reporting.

The CFO may personally review collections.

FP&A may maintain the operating forecast and help treasury when needed.

Those teams don’t necessarily need another complicated enterprise technology project.

They need reliable information without adding substantial administrative work.

If AI tools can simplify data integration, identify collection issues and help produce useful liquidity forecasts, smaller teams could benefit.

The financial value might come from reducing manual work.

But it could also come from improving decisions that previously required specialized analytical resources.

That’s a more interesting opportunity than simply automating another report.

It could make sophisticated cash management practices more accessible.

Of course, pricing, implementation requirements and data readiness will determine whether particular products are practical for smaller organizations.

FIS hasn’t established universal outcomes through its announcement.

Still, the direction is promising.

What FIS’s Launch Says About the Future of FP&A

I think we’re seeing a shift in how finance technology is being positioned.

For years, much of the conversation focused on faster reporting.

Faster closes.

Faster reconciliations.

Faster dashboards.

Those improvements matter.

But the next opportunity is helping finance teams respond to developments before they become financial problems.

Cash forecasting is a natural place to begin.

Because liquidity connects operating performance with actual financial capacity.

A company can report strong earnings and still face cash pressure.

It can also generate substantial cash while experiencing temporary accounting losses.

Understanding those differences requires more than looking at the income statement.

And it requires finance teams to connect operational activity with financial outcomes.

I think AI could make that connection more useful.

Not by replacing financial judgment.

By making relevant information available sooner.

That’s a worthwhile goal.

And one I would be happy to see more technology companies pursue.

The Cash Forecast Should Help Us Run the Business

I like the direction FIS is taking with Agentic Insights Suite.

The combination of multi-horizon cash forecasting, risk information and integrated financial data addresses a real business problem.

Finance teams need to understand where cash is going.

But they also need to understand what could change and what management can do about it.

That’s where the value lies.

I’d still want to test the platform.

I’d want evidence of forecasting performance.

I’d want to understand the implementation requirements and controls.

And I’d want to build a realistic financial business case.

But I wouldn’t dismiss the opportunity.

Better cash visibility can improve working capital, financing decisions and operating flexibility.

And those improvements can create real economic value.

The most useful application of AI in finance may not be the one that produces the most impressive presentation.

It may be the one that notices a $1.5 million customer payment is about to arrive late and gives the company three weeks to respond.

That’s not particularly glamorous.

It won’t inspire a keynote speech about the future of humanity.

But it might prevent an expensive borrowing decision.

And personally, I find that considerably more exciting than another dashboard.

Because revenue is wonderful.

Profit is important.

But when payroll arrives on Friday, the bank generally prefers cash.