How to automate your day as a journalists with AI

A journalists spends 30-50% of their time on repetitive tasks that don't need their brain. No judgment, no expertise, just clicks and copy-paste. Here's how to reclaim that time cleanly, without falling into classic over-automation traps.

Step 1 , Map what is repetitive

Over a typical week, keep a simple log. At the end of each task, note:

• Which task
• How long (to the minute)
• How many times in the week
• Judgment level required: 0 (mechanical) to 10 (pure expertise)

By end of week, you have your map. Every frequent (3+ times) and mechanical (judgment 0-3) task is an automation candidate.

Step 2 , Classify possible automations

Three categories, three approaches:

A. Simple routine (one click, copy-paste). Example: add each new client to your CRM, send a welcome email, create a Drive folder. Tool: Zapier or Make. Investment: 1-3 hours of setup, savings: 1-3 hours per week.

B. Text routine (standard writing). Example: responses to recurring client questions, template quotes, meeting notes. Tool: an LLM (Claude or ChatGPT) with a prompt template stored in Notion or TextExpander.

C. Decision routine (qualification, scoring). Example: evaluate if a prospect deserves a call, sort CVs, prioritize tickets. Tool: an LLM with a structured prompt + explicit rules + a human validation layer at the end.

Step 3 , The 5 highest-paying automations for journalists

1. Transcribe and search 10 hours of interviews in minutes.

Implementation: a prompt template stored in your favorite LLM, applied systematically. If the task is daily, you save 5-15 minutes per occurrence.

2. Cross-reference data and find angles in a dataset.

Implementation: a prompt template stored in your favorite LLM, applied systematically. If the task is daily, you save 5-15 minutes per occurrence.

3. Draft a first version to free up reporting time.

Implementation: a prompt template stored in your favorite LLM, applied systematically. If the task is daily, you save 5-15 minutes per occurrence.

4. Verify claims with traced sources.

Implementation: a prompt template stored in your favorite LLM, applied systematically. If the task is daily, you save 5-15 minutes per occurrence.

5. Spin one article into multiple formats (web, audio, video).

Implementation: a prompt template stored in your favorite LLM, applied systematically. If the task is daily, you save 5-15 minutes per occurrence.

Step 4 , Tools to stack intelligently

1. Descript ⭐ Recommended

For video or podcast journalists: lightning-fast text-based editing, subtitles, social formats.

Pricing : Free · $12 to $24/mo · Try free →

2. Fireflies.ai ⭐ Recommended

Transcribes interviews and surfaces key quotes. Instant search across an interview corpus. Essential for long investigations.

Pricing : Free · $18 to $39/mo · Try free →

3. Claude

For quickly turning raw notes into a story. Excellent editorial structure and sober journalistic style.

Pricing : Free · $18/mo (Pro) · $100/mo (Max) · Official site →

4. Perplexity

Source-cited search: verify claims fast and trace info to original sources.

Pricing : Free · $20/mo (Pro) · Official site →

5. ChatGPT

To analyze a dataset with code interpreter: stats, visualizations, anomalies. Powerful for data journalism.

Pricing : Free · $20/mo (Plus) · $200/mo (Pro) · Official site →

6. DeepL

To follow international press in languages you do not master, and translate foreign verbatims.

Pricing : Free · $9/mo (Pro) · Official site →

Step 5 , The standard automation system

Here's how effective journalists structure their stack:

Layer 1 (brain): you, making decisions and arbitrating.

Layer 2 (LLM): Claude or ChatGPT for producing, analyzing, summarizing.

Layer 3 (automation): Zapier or Make to connect tools.

Layer 4 (vertical tools): your CRM, accounting, storage.

Layers 2, 3, and 4 must talk to each other. That's where the magic happens.

Classic pitfalls to avoid

Pitfall 1: automating too early. Master the manual process before automating. Otherwise you automate chaos, which produces automated chaos, worse than manual chaos.

Pitfall 2: stacking 10 tools. Beginners want to test everything. Pros master 3 tools deeply. Start with 2-3, master, then add if a precise need emerges.

Pitfall 3: confusing AI with magic. AI accelerates what you know. It doesn't replace expertise. If you can't do something manually, AI will do it poorly and you won't see it.

Pitfall 4: automating client relationship. Welcome emails, OK. Cold reminders, OK. But the moment a client has a problem: you, in person, never AI alone. Otherwise you lose the client.

Pitfall 5: not measuring. Before and after each automation, measure actual time. Many automations are illusions that add complexity without real gain. ROI must be visible.

90-day rollout calendar

Weeks 1-2: mapping. No automation, just measurement.

Weeks 3-4: first prompt template. Pick THE most repetitive task and build a prompt that solves it. Use it for 2 weeks.

Weeks 5-8: add 2-3 more prompt templates. You start to see hours come back.

Weeks 9-12: first Zapier/Make automation. Connect two tools you use constantly (mail + CRM, or CRM + invoicing).

After 90 days, you've typically recovered 8-12 hours per week.

Going further

• Complete AI tools guide for journalists
• Ideal daily workflow
• Hard ROI numbers
• Complete pro stack

---

The right next step for a journalists

If you only test one tool this week, pick Descript. It is the one that comes up most often in community feedback for this profession. Free trial, no card.

Try Descript free →