Do I still need to learn manual journal entries if AI can do them?

Yes, more than ever. Understanding how journal entries flow, what controls catch errors, and why a JE is structured a certain way is the foundation that lets you spot when AI gets it wrong. The AI is the calculator; you're still the engineer who has to know what the answer should look like before pressing equals.

Should I learn Python as a junior accountant?

A working knowledge — enough to read code an AI agent writes for you, modify simple scripts, and run them — pays off enormously. You don't need to be a software engineer. Two weekends of effort gets you to useful, and from there it compounds for the rest of your career.

What Junior Accountants Should Learn in the AI Era

If you're a junior accountant in 2026, you're entering a profession in transition. The fundamentals matter more than ever; the tooling is changing fast. Here's how to spend your first three years to compound for a 30-year career.

Year 1 — Foundation that doesn't expire

The non-negotiables, in roughly the order to master them:

How to read a financial statement and trace any number to its source. Decades-stable skill.
Journal entries and accounting flow. AI will draft them, but you must know what they should look like to catch errors.
Internal controls reasoning. Why a control exists, what it's designed to prevent, what would go wrong if it didn't exist.
Excel: real Excel. INDEX/MATCH, XLOOKUP, structured references, pivot tables, conditional formatting, basic VBA. AI tools assume Excel fluency; you can't drive them well without it.
Working-paper discipline. Tickmarks, traceability, documentation standards. The audit reviewer doesn't care that AI produced it; the workpaper still has to defend the conclusion.
Communication. Email tone, memo structure, status updates, asking for help. AI can polish your writing; it can't make a non-communicator into a communicator.

Year 1 — AI fluency, in parallel

One AI chat tool (ChatGPT or Claude) used daily as a tutor. Learn the prompt patterns.
Microsoft Copilot for Excel. If your firm has it, learn it within month two.
One agent tool. Claude Code is the easiest entry point for non-developers. Build something — a CSV cleaner, a personal expense tracker — to learn the loop.
Your firm's sanctioned internal LLM. Most firms have one in 2026; learn what it can and can't do.

Year 2 — Specialization and depth

By year two you should be:

The team's default for "I have a weird Excel/data problem, who do I ask?" That visibility scales.
Reading code that AI agents produce for you. You don't need to write Python from scratch — you need to read and verify what an agent wrote.
Building one or two small repeatable workflows that you own. Every junior should leave year two with at least one "I built that" story.
Picking a specialization area: audit, tax, advisory, FP&A, technical accounting. The path branches faster in 2026 than in earlier eras.

Year 3 — Compounding skills

SQL. Reading data from real systems beats waiting for someone else to pull it. The ROI on basic SQL fluency is enormous.
Basic Python (read-and-modify level). Enough to take an AI-generated script and adjust it. Two weekends of effort.
One BI/visualization tool. Power BI or Tableau, depending on your firm's stack.
Project management. Year 3 is when juniors get pulled into running engagements. Without project management, your AI productivity gains get eaten by coordination overhead.
Industry depth. Pick an industry vertical and get fluent in its accounting nuances. Industry expertise compounds; generalism in 2030 is increasingly hard to monetize.

What to skip

Some things that used to be valuable are now low-leverage:

Manual data entry skills past basic competence. Useful to do once or twice for foundation; not worth scaling.
Memorizing every line of the IRC or every paragraph of the FASB Codification. AI handles lookup; your job is to know enough structure to know where to look.
Mastering the proprietary internal tools at your firm to the exclusion of transferable skills. Internal tools change; transferable skills compound. Spend 80% on the latter.

What never changes

Three things will define successful CPAs through 2030 and beyond, regardless of what AI does:

Judgment. The ability to weigh evidence, identify what matters, and make a defensible call.
Communication. Translating complex findings into clear language for non-experts.
Ethics and trust. Clients pay for trust; AI doesn't come with reputation.

Build those three first. Stack the AI tools on top. The combination is the most valuable accounting profile in the 2026-2030 market — and most of your peers won't put in the work on either side.

The 90-day starter plan

Days 1-30: Master your firm's ERP. Learn one AI chat tool well. Build your tickmark and workpaper discipline.
Days 31-60: Pick up Microsoft Copilot for Excel. Spend a weekend installing Claude Code and building a personal CSV cleaner. Practice writing client-facing memos.
Days 61-90: Learn enough SQL to query your firm's data warehouse for one common report. Volunteer for an engagement that's heavy on novel work — that's where junior visibility is highest.

Three months. The compounding starts in month four.