Day 25 of 28 · OpenClaw Challenge
Working with Data
⏱ 7 min
📊 Intermediate
Your agent can read documents and files. But what about data — numbers, trends, metrics, KPIs? The stuff that lives in spreadsheets and dashboards?
Today your agent becomes your data analyst.
Beyond reading — analyzing
On Day 12, you saw your agent read files. But data analysis goes further:
Reading = "What's in this spreadsheet?"
Analyzing = "What changed compared to last month? What's the trend? What should I worry about?"
Your agent can do both.
How did our sales compare this month vs last month?
Pulling your sales data...
Here's your month-over-month comparison:
📊 **Revenue**: $142,000 → $158,000 (+11.3% ↑)
📦 **Deals closed**: 23 → 28 (+5)
💰 **Average deal size**: $6,174 → $5,643 (-8.6% ↓)
🔄 **Pipeline**: $340,000 (up from $290,000)
🟢 **Good news**: Revenue and deal count both up
🟡 **Watch**: Average deal size dropped — you're closing more smaller deals
💡 **Suggestion**: Review the 5 new deals. If they're expansion accounts, the smaller size is fine. If they're new logos at lower ACV, might be worth investigating pricing.
Which of those 5 new deals were the smallest?
The 3 smallest new deals this month:
1. StartupXYZ — $2,100 (trial conversion, expected to expand)
2. DesignCo — $3,200 (new logo, single-team plan)
3. DataFlow Inc — $3,800 (new logo, annual contract)
StartupXYZ looks like a land-and-expand play. DesignCo and DataFlow are genuinely smaller deals. Want me to flag these for the sales team to review?
↻ Replay conversation
Knowledge Check
What did the agent do beyond just reading the sales spreadsheet?
A
It created a new spreadsheet
B
It only showed the raw numbers
C
It calculated trends, identified a concern (shrinking deal size), and gave an actionable recommendation
D
It made up the numbers to look impressive
The agent didn't just read — it analyzed. It calculated month-over-month changes, spotted a potential concern (declining deal size), and suggested a specific follow-up action. That's the jump from data reading to data analysis.
Automated reporting
Combine data analysis with automation, and your agent can generate reports on a schedule:
Create a weekly report from our support tickets. Send it to the team every Friday at 4pm.
Weekly support report configured ✓
Every Friday at 4pm, I'll:
1. Pull all support tickets from the past week
2. Calculate: total tickets, avg resolution time, top 3 issue categories
3. Compare against previous week's numbers
4. Flag any trends or outliers
5. Format as a clean report and email to the engineering team
Here's what this week's report would look like:
📊 **Support Report — Week of Feb 17**
**Volume**: 47 tickets (↓ from 52 last week)
**Avg resolution**: 4.2 hours (↑ from 3.8 — slightly slower)
**Top issues**:
1. Login timeout (12 tickets) — up 50% from last week ⚠️
2. API rate limits (8 tickets) — stable
3. Export formatting (6 tickets) — new this week
**🚩 Flag**: Login timeout tickets spiked. Recommend investigating — may be related to last week's auth update.
Looks good? I'll start sending these every Friday.
↻ Replay conversation
Knowledge Check
What makes an automated report more valuable than a manual data pull?
A
It runs consistently, compares against baselines, and flags anomalies — things humans often skip when doing manual reports
B
It costs less than hiring a data analyst
C
It's more accurate than manual counting
D
It uses fancier formatting
Automated reports do the comparison and flagging that humans skip when they're in a hurry. "Login timeouts up 50%" is the kind of insight that gets lost when someone is manually pulling numbers at 4:55pm on a Friday.
Ask questions in plain English
The best part: you don't need to know SQL, Excel formulas, or data analysis techniques. You ask in plain English:
"What's our customer churn rate this quarter?"
"Which product has the highest return rate?"
"Show me our top 10 customers by revenue."
"How does this quarter compare to the same quarter last year?"
The agent reads the data, does the math, and gives you the answer. No spreadsheet formulas, no pivot tables, no data exports.
Final Check
What's the biggest advantage of asking your agent about data versus analyzing a spreadsheet yourself?
A
The agent is always more accurate
B
You ask in natural language and get insights — no formulas, no pivot tables, no switching between tools
C
The agent can access data you can't
D
Spreadsheets are obsolete
The agent removes the technical barrier between you and your data. You don't need to know VLOOKUP or pivot tables. You ask a question in plain English, and the agent handles the analysis. The data is the same — the access is radically simpler.
📊
Day 25 Complete
"Your agent crunches numbers, spots trends, and builds reports. Data analysis without the spreadsheet wrestling."
Tomorrow — Day 26
Building Your Own Tools
What if the agent doesn't have a tool for what you need? Build your own.