How to Read the CPI Report with AI Before the Headlines Do

The June Consumer Price Index, or CPI, comes out this Tuesday. CPI is the government's monthly receipt for what everyday stuff costs, from groceries to rent to gas. Within minutes of it landing, you will see two versions of the same report. One headline says inflation is heating up. Another says the Fed can relax. Both are reading the same document. Both are just picking the number that fits their story.

You can skip the spin. The report is public, the math is simple, and a chatbot is genuinely good at reading a long table fast. What a chatbot is bad at is knowing which number matters, and whether one month means anything. That part stays on you. Here is the whole routine, with the real May 2026 numbers as an example, and a copy-paste prompt you can use the second the June data lands.

The four numbers to pull first

Before you open an AI, know what you are hunting for. The report is a wall of line items, and the signal hides in just four numbers.

Headline vs core. Headline CPI counts everything, food and gas included. The problem: food and gas prices jump around a lot from month to month. So economists also track core CPI, which is the same measure with food and energy taken out. Core moves slower, so it shows the real trend underneath. The Fed watches core, not headline. Mix these two up and you misread the whole report.

Month to month vs year to year. The year-over-year number compares this month to the same month a year ago. It is the big figure everyone quotes. The month-over-month number compares this month to just last month. It is the fresh signal, and it moves first. The yearly number can stay ugly for a while because old data is still baked in, so the monthly number tells you what is happening right now.

Now the real example. In the May 2026 report, released June 10, headline CPI rose 0.5% from the month before and 4.2% from a year earlier (BLS, May 2026 CPI). That 4.2% means something that cost $100 in May 2025 cost about$104.20 a year later. It was the hottest yearly reading since April 2023. Scary, if you stop there. But core CPI, with food and energy taken out, rose only 0.2% for the month and 2.9% for the year. The gap between 4.2% and 2.9% is the whole story. Gas and other energy prices jumped 3.9% in that one month, and that alone caused more than sixty percent of the headline rise. The steady trend underneath barely moved. Shelter (that means housing costs, mostly rent) makes up more than a third of the whole index. It rose 0.3% for the month and 3.4% for the year, still cooling down from its 2023 peak.

One report, three numbers, three different moods. Someone who only read "4.2%, highest since 2023" walked away with a genuinely misleading picture. Dodging that trap is what the rest of this post is about.

How to feed the release into a chatbot

The trick is to copy the actual report text into the AI instead of asking it from memory. Here is why that matters. A chatbot learns from data that stops at some cutoff date, so on July 14 it will not know the June numbers yet. Ask it "what was June CPI" and it may just invent a number that sounds right. Hand it the real text and that problem mostly goes away.

The BLS posts the summary at bls.gov/news.release/cpi.nr0.htm at 8:30 a.m. ET. Open it, select the summary section, copy it, and paste it into ChatGPT, Claude, or Perplexity. Some search-connected tools, like Perplexity, can fetch the report on their own, but I still paste the text so I know exactly what the AI read. On the last step, if a number is not in the text you pasted, do not trust it.

What to actually ask the AI

Ask questions that force the AI to quote the report back to you, not just describe the mood of it. Vague questions get vague answers. Here is the prompt I keep saved and reuse every month. Paste the report text, then this:

Below is the latest BLS CPI release. Using ONLY the numbers in this text,
answer in a short table:
1. Headline CPI: month-over-month and year-over-year.
2. Core CPI (less food and energy): month-over-month and year-over-year.
3. Shelter: month-over-month and year-over-year.
4. Which categories drove the monthly change the most?
Then in two sentences: did the CORE trend accelerate, hold, or cool
versus last month? Do not use any number that is not in the text below.
If a number is missing, say "not in release."

[paste the CPI summary text here]

The two lines that matter are "using ONLY the numbers in this text" and "do not use any number that is not in the text." They lock the AI to the source, so it is easy to spot when it wanders off. The last instruction, asking whether core sped up or cooled, is the one that keeps you from panicking over a loud headline.

A couple of good follow-ups once you have the table. "Was the monthly rise mostly gas, mostly housing, or spread across everything?" And "How does this month's core compare to the last three months?" That second one only works if you also paste the earlier reports, so the AI has old numbers to compare against.

Where AI gets CPI wrong

Most how-to posts skip this section. It is the one that actually saves you money. AI reads fast and lies with confidence, and a number-heavy report like CPI is exactly where that combination bites.

There is a quieter mistake too. Some numbers are "seasonally adjusted" and some are not. Seasonal adjustment just smooths out swings that happen every year, like heating bills each winter or shopping each December, so you can compare months fairly. The yearly headline number is usually reported without that adjustment, while the monthly change usually has it. An AI skimming fast will sometimes compare the wrong two, and then the numbers do not add up. If the AI's math looks off by a few tenths of a percent, this is often why. Ask it to say which basis each number is on.

My honest take, after doing this for a while: the AI is worth it for speed, and for yanking the shelter and category details out of a long table in seconds. It is not worth trusting as a judge of what the report means. Use it as a fast reading assistant, keep the judgment for yourself, and you get most of the upside with little of the risk.

How this connects to the Fed

CPI moves markets because of the Fed, and the link runs through interest rates. Quick background: the Fed is the U.S. central bank, and it sets one key interest rate that ripples out to your mortgage, your car loan, and the interest on your savings. That rate is called the federal funds rate. When core inflation stays high, the Fed has cover to keep that rate up. When core cools, the case for cutting it gets stronger. That is the chain a good read of the report is really tracing.

At its June 17, 2026 meeting, the Fed's rate-setting committee (the FOMC) held that rate in a range of 3.50% to 3.75% (CME Group, FedWatch). The next meeting is July 29, 2026. As of July 8, 2026, a tool called CME FedWatch, which turns bond-market bets into plain probabilities, showed traders pricing in roughly a 70% chance the Fed holds the rate steady again at that meeting rather than cutting it. Those odds move with every inflation and jobs report, so the June CPI on July 14 is one of the inputs that can shift them.

Here is the practical version. If June core inflation comes in soft, watch the odds tilt toward a rate cut. If core heats back up, expect the opposite. I am describing how the machine works, not predicting the number. None of this is a reason to trade on a single report. It is a reason to actually understand what you are reading while everyone else reacts to the headline.

FAQ

How do I read the CPI report myself with AI?

Pull the summary from the BLS release at bls.gov the moment it posts, paste the text into a chatbot, and ask it to extract headline and core CPI (both monthly and annual) plus shelter, using only the numbers in the text. Then check every figure it gives back against the release. The AI does the fast reading; you do the verifying.

What is the difference between headline CPI and core CPI?

Headline CPI includes everything, food and energy included, and those two swing sharply month to month. Core CPI strips them out to show the underlying trend. The Fed watches core because it is steadier. In May 2026, headline ran 4.2% year over year while core was 2.9%, because energy spiked and pulled the headline up.

Should I look at month-over-month or year-over-year CPI?

Both, for different reasons. Year over year is the widely quoted number and shows the twelve-month trend. Month over month is the freshest signal and moves first, so it tells you what changed most recently. A single monthly number, though, is a data point and not a trend on its own.

Can I trust ChatGPT to analyze the CPI release?

Only if you give it the actual release text and verify its output. Chatbots have a training cutoff and will not know the newest numbers, so asking from memory invites made-up figures. They also tend to over-weight a dramatic headline. Paste the source, box the prompt to that source, and confirm the numbers yourself.

How does the CPI report affect interest rates?

The Fed uses inflation trends, core CPI especially, to set the federal funds rate. Elevated core supports holding rates higher; cooling core strengthens the case for cuts. As of July 8, 2026, markets priced roughly a 70% chance the Fed holds at its July 29 meeting, and each CPI print can nudge those odds.

Sources

• U.S. Bureau of Labor Statistics, Consumer Price Index Summary (May 2026): https://www.bls.gov/news.release/cpi.nr0.htm
• BLS, CPI Release Schedule (June 2026 CPI released July 14, 2026): https://www.bls.gov/schedule/news_release/cpi.htm
• US Inflation Calculator, CPI Release Schedule: https://www.usinflationcalculator.com/inflation/consumer-price-index-release-schedule/
• CME Group, FedWatch Tool (federal funds rate and meeting odds): https://www.cmegroup.com/markets/interest-rates/cme-fedwatch-tool.html
• Data as of: May 2026 CPI (released June 10, 2026); Fed target rate and FedWatch odds as of July 8, 2026.