Most discussions about AI in finance are still surprisingly vague.

You'll hear people talk about AI agents, copilots, autonomous workflows, and digital workers.

But if you're leading Accounts Receivable, Credit, Shared Services, or Finance Operations, you're probably asking a simpler question:

What can I actually do with AI this week?

The good news is that you don't need to wait for fully autonomous finance departments to start benefiting from AI.

Many AR teams can unlock significant productivity gains today using tools such as ChatGPT, Claude, Copilot, and other AI assistants.

Let's look at some practical examples.

1. Turn Your AR Aging Into Executive Commentary

Most AR reports answer what happened.

Very few explain why it happened.

Instead of manually writing commentary every month, upload your aging report and ask AI:

This is one of the fastest ways to save time while improving reporting quality.

Pro Tip

Create a reusable prompt or Claude Skill so the commentary follows the same structure every month.

I'll have a free masterclass and showcase exactly how to do this, you can register here.

2. Create an AI Collection Prioritization Assistant

Many collection teams still work through customer lists manually.

AI can help rank collection priorities based on:

• Days overdue
• Invoice value
• Customer payment history
• Collection notes
• Open disputes

A simple Custom GPT, Gemini Gem, Copilot Agent or Claude Skill can generate a prioritized action list that helps collectors focus on the highest-impact accounts first.

You build it once, and then re-use when needed!

3. Build an AR Dashboard Without Coding

One of my favorite recent developments is Claude Artifacts.

Artifacts allow you to describe a dashboard in plain English and have AI generate a working application.

For example:

Within minutes, you'll often have a prototype you can refine further.

You don't need coding or developers. You just need a clear description of what you want.

You end up with a dashboard like this:

In the free masterclass I mentioned I will also showcase this, you can register here.

Pro Tip

Start by sketching the dashboard on paper first. AI performs dramatically better when given a clear structure.

4. Create Finance-Specific AI Skills

Most people use AI as a chatbot.

The more advanced users create Skills.

A Skill is essentially a reusable expert.

For example:

AR Analyst Skill

Every time you upload an aging report, the AI:

• Calculates collection KPIs
• Identifies high-risk accounts
• Flags unusual trends
• Creates management commentary
• Recommends actions

Instead of rewriting prompts repeatedly, the AI follows the process automatically.

Think of Skills as creating junior digital analysts for repetitive tasks.

5. Use AI to Draft Collection Communications

Many collection emails are repetitive.

AI can help you draft:

• Payment reminders
• Escalation notices
• Customer follow-ups
• Internal dispute summaries

For example:

The result is often 80% complete within seconds.

One thing that many people don't do is "personalize" the AI tools. In each one is in a different place but basically you can edit how AI will be responding. This helps you tailor de AI answers and avoid generic results.

6. Analyze Payment Patterns

AR teams often sit on years of payment history.

AI can help identify:

• Chronic late payers
• Seasonal payment behavior
• Customers likely to miss payment dates
• Emerging collection risks

These insights can improve forecasting and credit management decisions.

7. Automate AR Meeting Preparation

Before collection meetings, many teams spend hours preparing summaries.

AI can consolidate:

• Aging reports
• Collection notes
• CRM information
• Dispute status
• Recent customer interactions

Into a single briefing document.

For this I suggest you to prompt AI to get Python code to do the automation. Then you run the python-based automation in Google Colab (easiest to start) or Microsofr Visual Studio Code.

Where General AI Tools Reach Their Limits

Now for the important reality check.

Most AI tools are excellent at:

✓ Summarizing

✓ Explaining

✓ Drafting

✓ Building dashboards

✓ Generating insights

But Accounts Receivable isn't just an analysis problem.

It's also an execution problem.

Consider cash application.

A payment arrives.

The customer reference is incomplete.

Several invoices could match.

A credit note exists.

A short payment was taken.

A deduction is being disputed.

The ERP data is imperfect.

Suddenly the challenge isn't generating insight.

It's making a finance decision.

And making that decision in a way that's auditable, explainable, and trustworthy.

This is where specialized finance agents become interesting.

From AI Experiment to Autonomous AR

With Stacks, we're hosting a free live masterclass on how AI is evolving from assistant to operator inside Accounts Receivable workflows.

We'll cover:

The Finance Frameworks

• Why unapplied cash quietly distorts AR aging and working capital
• The three layers of the matching problem
• How leading finance teams think about cash application
• Common mistakes organizations make when applying AI to finance processes

Practical AI Demonstrations

• How ChatGPT, Claude, and Copilot can support AR activities
• How to build dashboards using Claude Artifacts
• How to create reusable AI Skills for finance teams
• What works today and what to avoid

First Live Look at Autonomous Cash Application

The most exciting part of the session will be a live demonstration from Albert at Stacks.

You'll see how AI can:

• How agents read the bank feed, the remittance advice, and the invoice to match a payment
• What an AI-suggested match looks like, and the confidence rationale behind it
• How exceptions get handled without holding up the close
• How cash application connects natively to bank reconciliation and the clos

If you've been experimenting with AI and wondering what the next stage looks like for finance operations, this session should give you a practical view of where the technology is heading.

Register here:

​https://stacks.ai/webinar-from-ai-experiment-to-autonomous-ar?utm_source=newsletter&utm_medium=email&utm_campaign=autonomous_ar_masterclass&utm_content=main_cta​

Date: June 30

Time: 5:00 PM CEST | 11:00 AM EST

I hope to see you there.

Christian Martinez