Excel's =COPILOT() is gone. Likely works in Excel and Google Sheets, and gives you numbers, not essays. Install for Excel →
⚡ FAST 5,000 rows in 1 second · Google Sheets + Excel

Stop reading every row. Sort by likely.

Type a question in plain English. Likely answers every cell with a probability from 0 to 1, 5,000 rows a second, so you can sort tickets, survey answers, leads and reviews instead of reading them.

2,000 rows a month freeNo credit cardRepeated cells are free

Support inbox · Sheet
fx
=LIKELY(A2:A9, "")
A · TextB · Likely

Anything with a yes/no, a label or a scale.

Speed

5,000 rows. 1 second.

Likely judges rows in big parallel batches instead of one at a time. Here's the same 5,000-row column, three ways, in real time.

⚡ Likely0.0 s
0 / 5,000 rows
🐢 A typical AI column0.0 s
0 / 5,000 rows
😩 Reading them yourself0.0 s
0 / 5,000 rows

Likely: measured on the live site with 5,000 unique rows, 0.8–1.2 s on warm connections, 1.6 s from a cold start. AI column: 1,000 rows a minute, the fastest bulk rate we found advertised by a popular AI add-on. By hand: 5 seconds a row.

The math

5,000 tickets. Which ones want a refund?

😩 Reading them yourself

  1. Open and read each ticket4 s × 5,000
  2. Decide, type a label1 s × 5,000
  3. Re-read the unclear ones+ coffee
Total6 h 57 min

😎 With Likely

  1. Type one formula20 s
  2. Likely judges 5,000 rows1 s
  3. Sort, check the yellow rows5 min
Total5 min 21 s

That's your afternoon back. Every week.

Why not a GPT column?

AI columns write essays. Likely gives you numbers.

A sentence in a cell can't be sorted, filtered or charted. A probability can.

A typical AI column

"Charged twice, refund one""Yes, this appears to be…"
"Billed after cancelling?""Possibly. The customer…"
"Export to CSV?""No."

Now parse it. And what does "possibly" mean across 10,000 rows?

Likely

"Charged twice, refund one"0.98
"Billed after cancelling?"0.27
"Export to CSV?"0.01

Sort descending. Act on the top. The unsure ones sit in the middle, honest about it.

How it works

Three steps. One formula.

🧩 Step 1

Add Likely

Install the add-on in Google Sheets or Excel and paste your free key.

💬 Step 2

Ask your question

=LIKELY(A2:A5000, "Is this a refund request?")

📊 Step 3

Sort, filter, act

They're numbers. Every spreadsheet feature you know works on them.

Also: LIKELY_CHOICE picks a label from your list · LIKELY_SCORE rates on your own scale · Bulk run for 200,000 rows

Used for

Any column you'd otherwise read by hand.

Every result below is a real answer from Likely.

🎧Support

Find the refund requests in this week's tickets.

"Cancel it and give me my money back"0.99
=LIKELY(B2:B, "Is the customer asking for a refund?")
💰Sales & RevOps

Rank leads by how ready they sound.

"Asked for a quote for 40 seats, starts next month"0.74
=LIKELY(D2:D, "Is this lead ready to buy soon?")
🔬Research

Code open-ended survey answers to your codeframe.

"Great app but way too expensive for a small team"price
=LIKELY_CHOICE(C2:C, "Main reason?", Codes!A:A)
📦E-commerce

Score every review and surface the angry ones.

"Arrived broken and support ignored me"Very negative
=LIKELY_SCORE(E2:E, "How negative?", "Positive, Neutral, Mildly negative, Very negative")
🛡️Community

Catch spam in comments before anyone reads it.

"WIN A FREE IPHONE, click the link in my bio!!!"0.96
=LIKELY(A2:A, "Is this comment spam?")
🧾Finance ops

Put transactions in your own chart of accounts.

"FIGMA.COM MONTHLY 15.00 USD"Software subscriptions
=LIKELY_CHOICE(F2:F, "Which account?", Accounts!A:A)
Pricing

Priced per row. Repeats are free.

A row is one non-empty cell judged. Blank cells, and text Likely already answered for the same question, cost nothing.

Free

$0 /mo

To try it on real work

  • 2,000 rows a month
  • Google Sheets and Excel
  • All three formulas
  • Bulk runs
Get free key

Team

$39 /mo

For the whole team's sheets

  • 2,000,000 rows a month
  • Everything in Pro
  • One key across your team
  • Cancel any time

Reading 10,000 rows by hand: 13 h 53 min. With Likely Pro: $12 and a coffee break.

FAQ

Good questions.

What does 0.87 actually mean?

It's the probability that your question is true for that cell. The model behind Likely is trained for calibration: across many rows, answers at 0.9 come true nine times in ten. Values near 0.5 mark text that could go either way. That's where a human looks.

Do you store my spreadsheet data?

No. Cell text is sent to be judged and is not kept. To make repeated cells free, we store a one-way fingerprint of each cell with its answer, never the text itself. Your email, key and monthly row count are the only other things we keep.

Why is it so fast and cheap?

Likely uses a model built for judgments, not writing. It returns a typed answer with its probability instead of generating sentences, batches rows into every call, and runs over a hundred calls in parallel. That's how 5,000 rows come back in 1 second. No essays: less waiting, smaller bill.

Formulas or bulk run: which should I use?

Formulas take up to 5,000 cells per range and update when your data changes. For bigger columns, use Bulk run in the Likely panel: it writes plain values, so nothing recalculates and nothing times out.

Can Likely write summaries or replies?

No, on purpose. Likely answers questions about your text: yes or no, which label, how much. That focus is what keeps it fast, cheap and sortable.

What if I hit my monthly limit?

Formulas show a clear message and cached answers keep working. Upgrade from the add-on in one click, or wait for next month.

Can I cancel?

Any time, from the Billing button in the add-on. You keep your plan until the end of the period you paid for.

Your next 10,000 rows, sorted before your coffee cools.

2,000 rows a month freeNo credit card