Analyst

Podcast 2 — In conversation

Speaker A: I've been offered a job title of Data Analyst and I'm wondering whether to push for Business Analyst instead. Is the distinction meaningful or is it just a label?

Speaker B: It's genuinely meaningful, though the boundary has blurred quite a bit in practice. A data analyst typically focuses on working with data directly — cleaning it, querying it, visualising it, and producing reports and dashboards. The outputs are usually charts, metrics, and statistical summaries that other people then act on. A business analyst tends to focus more on understanding business processes, identifying inefficiencies, gathering requirements from stakeholders, and specifying what systems or process changes are needed. The outputs are more often requirements documents, process maps, and functional specifications. That said, many organisations use both titles loosely, and plenty of business analysts do significant data work and vice versa.

Speaker A: And where does systems analyst fit in? I see that on job boards occasionally.

Speaker B: Systems analyst is an older title that was common before data analyst and business analyst became the standard terms. A systems analyst traditionally sits between the business and IT — they analyse how existing systems work, identify problems, and specify improvements, usually in terms of software or infrastructure requirements. The role is quite similar to business analyst in many organisations. It has declined in frequency as a job title, though the work still exists under other names. You are more likely to see it in public sector organisations or large traditional enterprises than in tech startups.

Speaker A: What about the phrase the analysts say or according to analysts — I see this constantly in financial news. Is that a specific type of analyst?

Speaker B: In financial journalism, analysts almost always refers to equity research analysts — the professionals at investment banks and research firms who cover specific industry sectors, publish research reports, set price targets for company shares, and issue buy, hold, or sell recommendations. When a news article says analysts expect earnings to rise, it is typically reporting the consensus of these equity research analysts rather than any other type. The implication is that these are people whose quantitative judgement on a specific sector carries professional weight — they have analysed the company's financials, its market position, and its outlook, and their views move markets.

Podcast 3 — Prompt engineering

Speaker B: Analyst in prompt engineering is one of the most productive role-assignment words in the entire vocabulary. When you instruct an AI to act as an analyst — or to produce output in the style and format that an analyst would produce — you shift its behaviour from generic description to structured examination, conclusion-drawing, and recommendation. The difference between asking an AI to describe data and asking it to analyse data as a senior analyst would is substantial. Let me demonstrate.

Speaker A: I've used act as a financial analyst in prompts before and got very good results. Is there a pattern to why that works?

Speaker B: Yes — role assignment activates a particular output register, structure, and depth. When you say act as a senior data analyst, the AI shifts to the vocabulary, format, and reasoning style of that professional role. Let's build through the scenarios. First — a business intelligence dashboard:

Speaker B: PROMPT:

Speaker A: The one-line interpretation note for each metric — that is the analyst's insight, not just the data. An analyst doesn't just report that revenue dropped three percent; they say a drop of this magnitude warrants investigating whether it correlates with a specific product category, region, or time window. That analytical framing is what makes the dashboard useful to a decision maker.

Speaker B: Now a requirements analysis prompt:

Speaker B: PROMPT:

Speaker A: The risk register is the detail that signals genuine analyst work. A developer describing a system would list features; an analyst documents risks, their likelihood, their impact, and mitigation actions. Adding that to the prompt tells the AI to produce a document a CTO would actually take seriously.

Speaker B: Now a data quality analysis prompt:

Speaker B: PROMPT:

Speaker A: The analyst framing turns a generic validation script into something that produces an audit-ready report. Pass or fail per check, counts, and sample rows — that is exactly the format a data governance analyst would present to a data steward before signing off on a migration.

Speaker B: Now a financial modelling prompt:

Speaker B: PROMPT:

Speaker A: The sensitivity table is the analyst's core tool — it shows not just the base case but how the conclusion changes as assumptions change. An analyst who presents only a single projection is not doing analysis; they are doing arithmetic. The sensitivity table transforms the tool from a calculator into an analytical instrument.

Speaker B: Now a security analysis prompt:

Speaker B: PROMPT:

Speaker A: Severity ratings — critical, high, medium — are standard analyst output in security work. Without that ranking, a development team with three critical findings and eight medium ones would not know where to start. The analyst's job is to prioritise, not just to enumerate.

Speaker B: Use analyst in prompts whenever you want structured examination, professional output format, conclusion-drawing, and prioritised recommendations rather than raw description. The word activates the AI's understanding of professional analytical standards — the registers, formats, and reasoning patterns that define the analyst role across data, business, security, and financial domains.