ANALYST TRACK · DAY 1 OF 5
Day 1 Day 2 Day 3 Day 4 Day 5

Stop summarizing. Start synthesizing.

A summary tells your boss what's in the documents. A synthesis tells them what to do about it. Claude can do the second. Most people only ask for the first.

Start Day 1 Day 2 →

The mistake most analysts make with Claude

You paste a 40-page board pack into Claude and ask "summarize this."

Claude returns 8 bullet points. Each bullet is accurate. The whole thing is useless.

Why? Because your boss didn't ask you for a summary. They asked for a recommendation. They asked "what's going on with vendor X?" or "should we approve the Q3 budget revision?" or "is this risk worth flagging to the board?"

A summary is a list of facts. A synthesis tells you what matters and what to do next. Most people use Claude as a faster reader. The useful version is using it as a thinking partner.

The shift isn't a new prompt. It's a new question. Stop asking Claude what's in the documents. Start asking Claude what to do about them.

Same input. Different output. Look at the difference.

Same document. Three different prompts. Three different outputs.

The document: a 12-page vendor security review (or any dense doc you didn't write — a Jira sprint review, a quarterly OKR doc, a customer feedback aggregation).

Prompt 1 — "Summarize this":
  • Vendor uses SOC 2 Type II, certified through Q4 2025
  • Three open audit findings from last assessment, categorized as medium severity
  • Last certification renewal completed Q4 2024
  • Encryption at rest: AES-256. Encryption in transit: TLS 1.3
  • Annual penetration test completed January 2026
  • Incident response plan updated November 2025
  • Vendor maintains 99.9% historical uptime
  • Sub-processor list reviewed and current

Eight true facts. Zero direction. You'd still have to read all 12 pages to make a decision.

Prompt 2 — The Decision Frame:

Decision: approve with conditions, or delay 30 days.

Real options:

  1. Approve with conditions — require closure of the three open audit findings within 60 days; escalate to security review board if missed.
  2. Delay 30 days — wait for findings to close before signing. Lower risk, but pushes the planned Q2 launch.
  3. Decline — no fundamental disqualifier, but the open findings suggest a vendor under-resourced on security. Defensible, probably overcautious.

Recommend option 1: leverage is highest pre-signature, findings are medium not critical, and 60 days aligns with our existing security review cadence.

Now you have something you can walk into the room with.

Prompt 3 — The Risk Pattern:

Biggest unstated risk: the vendor is likely under-resourced on security operations.

Specific signals from the document:

  • "Three open audit findings" — flagged in the previous assessment, still open in this one. They didn't close them in a full cycle.
  • "Incident response plan updated November 2025" — a recent update suggests they hadn't formalized one before, or are course-correcting.
  • The doc describes their security tooling but never names a CISO, security lead, or team. Larger vendors lead with this.

What's not being said directly: this vendor is probably understaffed on security headcount. That doesn't disqualify them — but the risk shifts to your team to monitor, not theirs.

Same 12 pages. The synthesis tells you what to decide. The risk read tells you what nobody wrote down.

Note on real output: Your output will look different. Claude will hedge more ("Based on the document, it appears..."), and you'll often need one or two follow-up prompts to sharpen the response. The shape of the output is what matters, not the specific words.

Two prompts that replace 90% of "summarize this"

Copy them. Use them this week.

1. The Decision Frame

Based on these documents, what's the decision I need to make, and what are my real options?

Use when: you have a stack of inputs and unclear next steps.

2. The Risk Pattern

What's the biggest risk in this material that isn't being said directly? Quote the specific lines that hint at it.

Use when: stakeholders are talking around something. This surfaces what's implied but not stated.

These two prompts pull on different muscles. The Decision Frame turns information into action. The Risk Pattern turns silence into signal. Most analyst work needs both.

Today's exercise (15 minutes)

Pick a vendor SOW, a Jira sprint report, or a recent meeting transcript. Something you'd usually skim through Claude as a "summarize this."

Don't summarize it.

Run both prompts back-to-back.

Based on this [document], what's the decision I need to make, and what are my real options?

Then:

What's the biggest risk in this material that isn't being said directly? Quote the specific lines that hint at it.

Compare both outputs to what a "summarize this" prompt would have given you.

The Decision Frame points at action. The Risk Pattern points at what's hiding. Together they replace the summary with something you can act on.

Tomorrow

Day 2 is where this gets compounding. You'll set up a Status Vault — a Claude Project that holds your stakeholders, your recurring formats, and your source-of-truth docs. Once it's set up, outputs improve because Claude already has your context.

Tomorrow, you stop re-explaining yourself every session.

Analyst Track Progression

Your 5-day path to a weekly analyst rhythm.

Day 1 → Day 2 → Day 3 → Day 4 → Day 5