The Analyst Who Can Explain the Answer

UBS has given aspiring bankers a fresh reason to revisit their financial analyst skills. Knowing the formulas is only part of the preparation.

On September 6, the Financial Times reported that candidates for the bank's 2027 graduate and internship intake in Global Banking and Markets would need to demonstrate AI proficiency. The reported emphasis: using it to improve outcomes and efficiency.

That creates a much better question than 'Which chatbot should I learn?' Can you put a piece of work on the table and explain exactly why someone should trust it? Let's try.

Imagine this assignment. You are reviewing a short company update before it goes to a senior colleague. The company and every figure below are hypothetical, created for this exercise.

That is an intentionally flawed sample summary. It is fluent, compact, and wrong in three useful ways. Before reading on, identify the problems. You need a calculator and the six numbers above. No AI subscription required.

First, check the denominator. Revenue grew by $20 million on a $100 million starting base. That is 20%, not 25%. Ask which period each input belongs to before accepting any calculated growth rate.

Second, distinguish growth from profitability. Net margin fell from 10% to 7.5%: $10 million divided by $100 million, versus $9 million divided by $120 million. Sales grew, but each dollar of sales produced less profit. The summary reversed the economic message.

Third, follow the cash. Using operating cash flow less capital expenditure, free cash flow was negative $2 million. Current operations did not cover that year's capital spending. Existing cash or financing could cover the gap, but positive accounting earnings do not establish that the programme funds itself.

The revised conclusion is more useful: sales expanded, margins weakened, and the investment programme exceeded operating cash generation. Now a decision-maker knows what to investigate.

Here is the framework I'd use to organise the review: Trace, Reconcile, Defend. It works whether the first draft came from AI, a spreadsheet, or another person.

For example, the margin decline deserves investigation. It does not tell us, by itself, whether pricing weakened, input costs increased, or the company deliberately spent more to expand. Identifying the question is progress; inventing its answer is not.

That combination is what our Financial Modeling & Valuation (Core + AI) program is built around: core Excel modeling and valuation, followed by analysis and modeling with ChatGPT and Claude. Learn how the model works, then practise using AI to build it and check the output.

Build one small, explainable example. Use a public annual report. Set a narrow task, such as comparing two years of revenue and profitability. Save the source table, your instructions, the draft output, and the corrections.

Then practise a plain-English explanation: 'I used the tool to organise the disclosure. I checked the periods and recalculated the margins. I changed the conclusion because the draft confused higher sales with better profitability.' Only make that claim after actually doing the work.

An unfamiliar formula is a warning to slow down. If you cannot explain working capital, a generated cash-flow forecast can look plausible while hiding a serious mistake. If you cannot explain the valuation assumptions, a precise target price can create false confidence.

Practise both modes: complete a small calculation yourself, then use AI and compare. Note where it saved effort, where you intervened, and which checks remained necessary. Use public or employer-approved material and follow the organisation's rules for tools and data.

For recruitment, follow the stated assessment rules. Showing that you can use AI responsibly does not imply permission to use it during every test or interview.

Our guide to seven essential financial analyst skills connects this exercise to Excel, accounting, financial analysis, valuation, and communication. Pick the weakest link in your own work and start there.

Reply and tell me which error you spotted first: growth, margin, or cash flow. I'd also like to hear which part you found hardest to explain.

Best regards,

P.S. Want to move from checking a short summary to building and reviewing a full model? Our FMV Core + AI program connects the finance fundamentals with practical AI workflows. Take a look through the curriculum and see whether it fits your next step. The newsletter stays free either way.

Home furniture brand Blu Dot blew up on CTV with help from Roku Ads Manager. Here’s how:

After a test campaign reached 211,000 households and achieved 1,010% ROAS, the brand went all in to promote its annual sales event. It removed age and income constraints to expand reach and shifted budget to custom audiences and retargeting, where intent was strongest.

The results speak for themselves. As Blu Dot increased their investment by 10x, ROAS jumped to 2,308% and more page-view conversions surpassed 50,000.

“For CTV campaigns, Roku has been a top performer,” said Claire Folkestad, Paid Media Strategist, Blu Dot. “Comping to our other platforms, we have seen really strong ROAS… and highly efficient CPMs, lower than any other CTV partner we've worked with.”

Using Roku Ads Manager, the campaign moved from a pilot to a permanent performance engine for the brand.

Learn More