Session recording
Alastair MatchettFounder and Managing Director, Financial Edge Training
“This is a view inside the world's biggest banks and what they're seeing in their junior professionals. It's unbelievable.” Ian Schnoor, Executive Director, Financial Modeling Institute, during the 24-Hour Global Financial Modeling Conference
The recording of Alastair's session is included with FMI Foundations and the Advanced Financial Modeler (AFM) Program, along with every other session from the conference.
But Alastair's super prompts and model files are free. Download below.
The two super prompts Alastair ran live, with the Coca-Cola (KO) workbooks they produced. The WACC and DCF prompt works for any public or private company; the 3-statement prompt assumes Financial Edge's Felix data connector and its model template.
Sourcing rules, a four-tab layout, peer betas, stub-period and mid-year discounting, and 17 checks to run before delivery.
Rules learned over a full Claude build: verify every historical against the filing, a formula-driven balance check, no plugs.
Coca-Cola valued at 10 September 2026 at a 7.6% WACC: $78.37 a share by Gordon growth, $88.74 by exit multiple.
A five-year KO model, FY26 to FY28 anchored to consensus, a zero balance check every year and a full Sources tab.
“Do not fabricate historical financial statements merely to make the model appear complete.”
Alastair expects AI to do most of the model building by 2028. His argument is that this raises the bar on fundamentals rather than lowering it. An analyst who cannot build a model cannot write a good prompt, check what comes back or defend it in the room.
Write the spec: structure, outputs, template, skills. A basic prompt gets garbage out. A detailed one gets a much higher quality result.
AI tools are moving to per-token pricing. Templates and saved skills give the AI a known structure: better output, fewer tokens.
These models are probabilistic. Formula errors, hallucinated links and commentary that reads well but turns out to be rubbish are caught only by someone who knows the right answer.
AI takes over the building. It does not take over the knowing. At one firm, the analysts using the most tokens were not the strongest performers: without the fundamentals, more AI produced more generic output.
The first prompt is rarely the last. Keep correcting the model in the chat until it is right, then ask the tool to write a super prompt based on that chat. If you will use it again, save it as a skill, a stored prompt you call with a forward slash, so the next build starts where the last one finished.
Ask the tool to walk you through its steps in plain language. Then set up a project whose instructions say you are a senior managing director covering the sector, rigorously checking your analyst's work. Run every output through it to find errors, size the valuation impact, sensitize the model and draft the commentary. You still have to defend the result yourself.
Alastair launched four builds at the start of the session: in Rogo, a basic DCF prompt and a longer super prompt for Coca-Cola; in Claude, a basic and a refined prompt for a 3-statement operating model, using Financial Edge's Felix data connector, a template and saved skills. When he ran the Claude build with a template earlier, it took 12 minutes and 191,000 tokens.
The tools still make mistakes. In one Rogo run, the implied terminal value multiple came out wrong: the Gordon growth value sits at mid-year, the multiple at year end, and the discounting did not adjust. His reusable WACC and DCF prompt spells out the fix: roll the value forward half a year first.
Each quote comes as an image for LinkedIn, dark or light, with the post already written. Pick one, download the image, and post it in three steps.
“I do predict that AI will do most of the modeling, but it still means you need to learn the fundamentals.”
“When you get AI to create a narrative about a model, it can do a very good, compelling job, but often it can be rubbish.”
“If you put a very basic prompt into the tool, you will often get garbage out.”
“The quality of the output was very, very strongly linked to the quality of the input prompt.”
Two ways to carry the conference forward: learn how a model is put together, or prove you can build one from scratch. Both FMI programs include every conference video and the presentation materials.
Learn core financial modeling concepts at your own pace. Access all conference videos and presentation materials.
AI can generate a model in seconds, but it can't tell you if the logic holds. The AFM proves you can build trusted financial models. Access learning resources and the full conference video library.
Pick a dark or light background, then save the image. It is a 1080 by 1080 square, which LinkedIn shows in full.
The post below opens in LinkedIn already written, and goes on your clipboard too. Change it here first if you like.
In the composer, choose the photo icon and add the image you downloaded. Type @ to tag Alastair Matchett and Financial Modeling Institute, then post.
