Future of Finance Lab Session 11 – AI Slop, Excel’s Future, and what the benchmarks actually show to help you boost your career to the next level

Background

Session 11 of the Future of Finance Lab cuts through the noise, with data, benchmarks, and no agenda to sell you a tool.

Paul Barnhurst, known across the finance community as The FP&A Guy, joined Lance Rubin for a session that had been building for months.

Both had watched the same playbook repeat itself across LinkedIn: provocative hooks, straw-man arguments about Excel, and calls to action dressed as insight.

Session 11 was the response.

A structured look at what AI can actually do in financial modelling, backed by real evaluation data, and a clear articulation of what finance professionals need to do about it.

Key Takeaways — Session 11

• AI models score <50% on financial models without domain skills — and 64–77% with EXL Cloud skills
• Three types of AI slop exist: vendor slop, creator slop, and adjacent exec slop — learn to identify each
• The five validation questions every AI-built financial model must pass before you put your name to it
• The Orchestrator Modeller role is replacing the traditional finance builder — and how to make the shift

Future of Finance Lab Nugget

Adding human-designed skills to AI raised model quality from 26% to 64%.

That’s a 38-percentage-point lift, yet the human-built model still scored 93%. 

That gap is not a prompt engineering problem.

It is a domain expertise problem. You can only write skills that prevent AI from making mistakes if you already know what a mistake looks like.

Foundations are not becoming less important. They are becoming the only reliable differentiator. This is what we test in the Future of Finance Lab.

What Did AI Benchmarks Actually Show in 2026?

What Are the Three Types of AI Slop in Finance?

Lance and Paul have developed a working taxonomy for the content polluting finance feeds. There are three distinct types, each with its own playbook.

Vendor slop is pitch dressed as prediction. A founder declares their AI tool will replace an entire category, and the bold claim is really a call to action. Shortcut’s “superhuman Excel agent” is the clearest recent example — it went viral with 9,000 comments and a YouTube video titled Rest in Peace Excel. Within a month, the company had quietly pivoted 80–90% of its focus to an Excel add-in. The market found them out faster than the algorithm did.

Creator slop is newsletter bait. “Build a five-year model in minutes” drives subscribers and impressions, and the urgency and novelty are the point — not the output. The hook goes to identity and anxiety: if you believe Excel is dying and AI can do it all, you feel behind, you click, you subscribe. The call to action is always there. The quality of what the AI actually built is never shown in full.

Adjacent exec slop is perhaps the most damaging. Hot takes from people outside the domain — HR founders declaring Excel dead, software executives opining on accounting, non-modellers claiming nine AI agents can replace a CFO. The authority is borrowed from their role, not earned from the domain. The question to ask: has this person actually done the work they’re claiming AI can replace?

The playbook is consistent: provocative hook, a straw man (12 tabs, deeply nested formulas), magical AI solution, a prediction set 12–14 months out, and a call to action. Once you recognise it, you start seeing it everywhere. The antidote is to ask who posted it, what domain expertise they actually have, and what they’re ultimately trying to get you to do.

The Real Benchmarks: What AI Delivers in Financial Modelling

Lance ran a structured evaluation using a complex SaaS model — equity funding rounds, convertible notes, cap table, term debt — as the test case. Five models were assessed across nine criteria including structure, formula logic, scenario engine, validation checks, and accuracy.

Results: Microsoft Copilot (GPT 5.5, no skills) and Claude Opus 4.8 cold both came in well under 50%. Claude Opus 4.8 with EXL Cloud skills: 64%.

Fable 5 with skills: 77%.

Human-built EXL Cloud model: 93%.

Craig Hatmaker’s 5G Lambda-optimised build: 85%.

Independent research supports the pattern.

Wall Street Prep’s 2026 benchmark scored Shortcut at 5.9 out of 10 against a junior analyst’s 6.4, and all four AI tools tested failed to resolve circularity.

Every tool used hardcoded plugs. MIT research found 95% of enterprise generative AI produces zero measurable P&L impact.

Microsoft’s own Copilot documentation states it should not be used for numeric calculations. The vendors are telling you directly what the tools cannot do.

Craig was a previous guest on the Future of Finance Lab and long been an innovator in this space.

Why Do AI Models Fail on Balance Sheets?

Across every AI tool tested, the P&L came out reasonably well.

The balance sheet was a different story. Accuracy ranged from 20–30% across AI-generated models, an 80% error rate.

If the balance sheet is off by 80%, you can’t be confident in your P&L or cash flow either.

The problem is structural.

When asked to balance a balance sheet, AI does one of two things: it creates a hardcoded plug, or it introduces a circular reference. Both are unacceptable.

The AI doesn’t inherently understand why, it needs to be told through skills.

Without that context, it guesses. On a complex balance sheet with convertible notes, equity waterfall mechanics, and term debt amortisation schedules, the guesses fall apart.

Paul raised a related point: even when AI gets the numbers right, the formula construction is often illogical. Revolvers calculated inline in the income statement rather than on a separate schedule. Scenario engines that rebuild the entire model three times rather than using a single toggle, tripling validation work rather than reducing it.

That was ChatGPT’s approach when given the FMI Henderson case.

Polished does not mean correct. That is the central warning from this session.

How Should Finance Professionals Respond to AI in 2026?

Skills: The Bridge Between AI Hype and Professional Results

• The 38-percentage-point lift from adding EXL Cloud skills, 26% cold to 64% augmented, demonstrates the real lever.
• Out-of-the-box AI and skills-augmented AI are not the same tool. The difference comes from human expertise encoded into the AI’s instructions.
• Skills are human-readable, markdown-format instructions that transfer professional judgement to the AI. They specify: no hardcoded actuals, no circular references, always include a validation sheet, always include an audit log, the balance sheet must reconcile. These are things a professional modeller knows instinctively. The AI does not, until it’s been told. EXL Cloud now has over 170 skills in production, with some available on GitHub.
• The catch: you can only write skills that prevent mistakes if you already know what a mistake looks like. You can’t instruct AI not to build a broken revolver schedule if you’ve never built a correct one.
• The finance professionals who extract the most value from AI are those with deep domain expertise who know when the output is wrong. AI is a magnifier. It magnifies strong skills and weak ones equally. Skills narrow the gap, but they don’t close it without underlying human expertise.

What Is the Orchestrator Modeller Role?

• The finance professional’s role in the AI era is not disappearing. It is changing from builder to orchestrator.
• You define the purpose. AI drafts the structure.
• You review and challenge. AI accelerates the build.
• You validate each component and own the output, the assumptions, the formulas, and the logic behind every number.
• Five questions before trusting any AI-generated model:
  • Does the balance sheet balance?
  • Are actuals hardcoded?
  • Is circularity present and how is it handled?
  • Are there validation checks?
  • Can the builder explain the logic behind every formula?
• If any can’t be answered confidently, the model is not ready to inform a decision, and putting your name on it is a career risk.
• Paul added a point worth sitting with: many finance professionals were never taught good modelling design principles during their corporate careers.
• There is a real temptation to accept AI output that looks more structured than what they’d have built themselves, without recognising that polished structure and reliable accuracy are completely separate things.
• Use the gap between where your skills are and where they should be as a reason to close it, then use AI from a position of genuine expertise.

Important Quotes from the Session

• On AI slop:
  • “An HR founder declaring Excel dead, that’s a problem. Yet the Big Four are going big on AI, and they are not hiring less. They’re hiring more people to do this work.” — Lance
  • “Someone with no finance experience shared their nine agents to replace CFOs built with Claude. And I’m supposed to trust you?” — Paul

• On the Shortcut story:
  • “Within a month of that video titled ‘Rest in Peace Excel’, they completely switched to focusing on their add-in inside Excel. Where did he end up back to?” — Paul
  • “If you’re not building something inside Excel, where people work, where the real work gets done, you’re not even a starter.” — Lance

• On AI accuracy:
  • “The tool literally told me: it’s okay if the balance sheet doesn’t balance. A 0.9% variance is acceptable. I don’t know what world you live in.” — Paul
  • “All four AI tools failed to resolve circularity. Every tool uses hardcoded plugs.” — Lance (on the Wall Street Prep benchmark).

• On skills:
  • “Without the skill, with the skill — a 38-point difference. But still 64%. If I got 64% on my college exams, that’s not looking too good. What if that 64% means you can do twice as much because you just have to fix the 36?” — Paul
  • “You can only build skills if you actually understand what good looks like.” — Lance

• On foundations:
  • “Those who get the most out of AI are those that understand the domain very well. AI is a magnifier. If you know what you’re doing, it magnifies that. If you don’t know what you’re doing, it magnifies that.” — Paul
  • “AI is getting upskilled. What are we doing? Where are we meeting AI?” — Lance

• On the probabilistic nature of AI:
  • “The nature of the technology is a probabilistic engine, it’s not suddenly going to become deterministic. So we’re always going to have a human in the loop.” — Lance

• On career risk:
  • “Two big banks came to him saying: we’re struggling to find people out of school because they all think AI can do it, and they don’t have the skills.” — Paul
  • “If you take a model to someone and say ‘what’s the assumption?’ and your answer is ‘I don’t know, the AI built it’, that is career limiting.” — Lance

• On closing:
  • “No technology has destroyed jobs. It has destroyed industries. It has destroyed specific jobs. But they don’t destroy economies, they just change economies.” — Paul
  • “The genie is not going back in the bottle. It’s like publishing a book. Once the knowledge is out, the knowledge is out.” — Lance

The Future of Finance Lab, a Journey of Learning and Adaptation

This session encapsulated the core message of the entire Future of Finance Lab series: while technology and AI are transforming the finance function, the human element—curiosity, adaptability, and best practice, remains irreplaceable. The series has equipped participants with frameworks, tools, and mindsets to navigate ongoing change, emphasizing that continuous learning and critical thinking are the keys to success.

Call to Action

Ready to future-proof your finance skills?

Explore the full Future of Finance Lab series for in-depth sessions, practical resources, and expert insights. Whether you’re a seasoned modeler or just starting out, there’s always more to learn—and the journey is just beginning.

Check out the recordings, download the workbooks, and join the conversation as the world of finance continues to evolvecussions, and resource sharing, this ongoing and free series remains a must-attend for anyone looking to future-proof their finance skills.

Looking Ahead

The Future of Finance Lab is more than a webinar—it’s an interactive community dedicated to learning, sharing, and growing together. We invite you to join our next session, connect with peers, and continue building the skills that will define the future of finance.

For resources, recordings, and more, visit our Knowledge Hub and stay tuned for further updates!

The session was also recorded and is available to be viewed on YouTube.

---

---

Want to discuss how AI can transform your finance function? Get in touch with Model Citizn to learn more about our approach to AI-powered financial modelling and strategic implementation.

Want to take action? You can by clicking here

We launched our Future of Finance Lab series of FREE monthly lightening sessions hosted on Maven to unpack all the hype and give solid foundations to finance professionals. Join our Friday Future of Finance Lab by registering for the lightening sessions on Maven.

Friday’s Future of Finance Lab

Join our community where finance professionals share insights, solve real challenges, and stay ahead of the AI revolution (max 50 attendees).

Key Benefits:

✅ Live Problem Solving – Bring work challenges

✅ Expert Insights – Practitioners not trainers

✅ AI Integration – Leverage AI in finance

✅ Networking – like-minded professionals

✅ Zero Cost – Completely free, always

What You’ll Get:

✅ – Monthly Focus Areas in the following areas

✅ – Financial Modeling Deep Dives

✅ – Data Analysis & Power BI Techniques

✅ – AI Applications in Finance

✅ – Open Forum & Case Studies Exclusive Resources

✅ – Meeting recordings and AI-generated summaries

✅ – Templates and tools shared during sessions

✅ – Resource library access

Who Should Attend:

✅ – Finance professionals wanting to upskill

✅ – Data analysts working with financial data

✅ – Anyone curious about AI in finance

✅ – Business leaders seeking data-driven insights

Buckle up and look forward to seeing you there at the Future of Finance Lab!

AI Accounting Finance #CFO #FinancialModeling #AIAdoption #FutureOfWork

Author

Lance Rubin is CEO and Founder of Model Citizn and EXL Cloud. He is a Chartered Accountant, FMI Certified Trainer, co-author of the CA ANZ Financial Modelling Study Guide, and creator of the AI-Powered Accountant programme on Maven, currently live at Level 1 with Level 2 in design, and author of the companion book “AI-Powered Accounting with Excel and Power BI” (Packt). He has worked at PwC, KPMG, Investec, and NAB, and is based in Melbourne, Australia.