Illustrative study with synthetic respondents — not the company's data. Pricing teardown · Microsoft 365 · January 2025

Microsoft put Copilot in Microsoft 365 Personal and Family and raised prices by $3. Who paid more?

On January 16, 2025, Microsoft raised its consumer Microsoft 365 plans by $3 a month in the US, the first rise since launch, and included a monthly allowance of Copilot AI credits. Personal went up 43%. Subscribers who didn't want AI could ask for a Classic plan at the old price. This teardown looks at the decision behind it, and how you'd test it.

Public sources only, each citedPrices in USD per subscription per monthChecked October 3, 2026
In short
  • Personal went from $6.99 to $9.99 a month and Family from $9.99 to $12.99, each with about 60 AI credits a month for Copilot in the apps.1,2
  • Unlike a business bundle, few people were paying for AI here: Copilot Pro cost $20 a month on top.2 So the rise fell on almost everyone.
  • Microsoft offered Classic plans without AI at the old price, for a limited time.1 The test: does the rise earn more than it loses to people who don't want AI?
Situation

AI was a separate $20 subscription

Before January 2025, Microsoft 365 Personal (one person) and Family (up to six) included the desktop Office apps and 1 TB of storage per person, but no Copilot in the apps. AI in Word, Excel and Outlook came with Copilot Pro, at $20 a month.2

Microsoft 365 consumer plans before January 16, 2025 (US)

PlanBilled monthlyBilled yearly
Personal$6.99$69.99
Family$9.99$99.99
Copilot Pro$20—

Per subscription. Microsoft 365 annual prices as reported by Computerworld ($70 and $100, rounded there); monthly prices are $3 below the new ones Microsoft announced.1,2

Task

Keep AI as a $20 extra, or put some of it in every plan?

Microsoft gave its reason in the announcement: the price reflects the extensive subscription benefits that we've added over the past 12 years and funds new ones.1 Microsoft has not published how many consumers paid for Copilot Pro.

So the pricing decision was this: a small rise for every subscriber, with a limited AI allowance, against a $20 extra that few people bought. And what happens to the subscribers who don't want AI?

Before deciding: what you'd want to know

Four questions a pricing study can answer

  1. Which features make a home office subscription worth paying for: the AI or the rest?MaxDiff →
  2. Should AI be sold as a $20 extra, or included with a usage allowance for $3 more?Conjoint →
  3. What is the acceptable price range for Personal with AI credits included?Van Westendorp →
  4. At what price does Personal with AI credits bring in the most revenue?Gabor-Granger →

The order runs from the offer to the price: what buyers value, how to package it, the range they accept, then the exact price. Each question is asked separately of the two groups the change affects differently: people who already pay for an AI assistant, and people who don't. The study, run in Albi, is at the end.

Action

Both plans went up $3, with AI credits included

On January 16, 2025, Microsoft announced that Copilot was included in Microsoft 365 Personal and Family, and that US prices would rise by $3 a month. Existing subscribers saw the new price at their next renewal.1

PlanBeforeFrom January 16, 2025
Personal$6.99 · no AI in the apps$9.99 · AI credits included
Family$9.99 · no AI in the apps$12.99 · AI credits, owner only
Personal or Family Classic—Old price · no AI, limited time
Copilot Pro$20$20 · for more AI use

Billed monthly, US. About 60 AI credits a month2; in Family, only the subscription owner gets the AI features.3 Classic plans keep the old features without the new AI ones.1

The same $3 is a bigger rise for Personal

$0$5$10$15Personal · before$6.99Personal · after$9.99Family · before$9.99Family · after$12.99
Per subscription per month, billed monthly, US list prices1,2: +43% for Personal, +30% for Family. Our arithmetic; it leaves out discounts and other countries.
Result

What is public about how it went

Microsoft has not published how many subscribers moved to Classic, cancelled, or used their AI credits. What is on the record:

  • On October 1, 2025, Microsoft launched Microsoft 365 Premium at $19.99 a month, combining Family with Copilot Pro's features, and stopped selling Copilot Pro on its own.4 That is a separate pricing decision, not covered here.

What isn't public: how many subscribers took Classic, and whether the AI credits changed who renews. This teardown makes no claim about either.


The study

From the four questions to the decision

Illustrative study with synthetic respondents — not the company's data.

We ran the four questions above as one Albi survey. The respondents are US adults who pay for a home office-apps subscription for themselves or their household: 150 who already pay for an AI assistant, and 150 who don't.

These results come from synthetic respondents: AI-modelled buyer profiles, not people. Use them to shape the question, then confirm with real respondents. Run in Albi on October 3, 2026, with 300 synthetic respondents generated by Gemini 3.5 Flash-Lite. Every chart below meets the minimum sample Albi sets for its method.

1

Which features make a home office subscription worth paying for: the AI or the rest?

MaxDiff · all 300
10 = average Full desktop office apps23.1 Advanced file, email security12.4 AI summaries, drafts in email11.9 AI help with spreadsheets11.9 1 TB cloud storage10.3 AI drafting in documents9.8 Share with up to 5 others7.4 AI presentation creation6.1 Ad-free email5.4 AI image creation1.7
Preference score; the 10 features sum to 100. Purple: AI features. Grey: the rest of the subscription. Names shortened; the survey described each feature in full. Synthetic respondents, n = 300.

Result. The desktop apps came first by a wide margin (23.1, more than twice the average). Advanced security came second (12.4), with AI in email and AI help with spreadsheets just behind (11.9 each). The five AI features together hold 41 of the 100 points, below their half of the list, and AI image creation came last. The groups split sharply: people who already pay for AI put AI drafting far ahead of everything; people who don't put the desktop apps first and AI drafting near the bottom.

For the decision: the subscription is bought for the apps. AI adds value for people who already use it, and almost none for the rest, who pay the same $3.

2

Should AI be sold as a $20 extra, or included with a usage allowance for $3 more?

Conjoint · all 300 · 5 prices × 3 AI levels × 2 plan sizes

Old packaging vs new: who buys what, and what it earns

Before January 16, 2025

AI sold separately as Copilot Pro

Personal$6.99
1 person, 1 TB
Family$9.99
up to 6 people
Copilot Pro+$20
AI in the apps
Share of buyers
Share of revenue
BuyersRevenue
No plan bought (vs. these options)4%—
Personal70%61%
Family26%33%
Copilot Pro2.5% take it6%
Monthly revenue per 100 buyers$802
From January 16, 2025

AI credits included, price up $3

Personal$9.99
AI credits included
Family$12.99
AI credits, owner only
Copilot Pro+$20
higher AI limits
Share of buyers
Share of revenue
BuyersRevenue
No plan bought (vs. these options)12%—
Personal76%83%
Family12%17%
Copilot Pro<1% take it0%
Monthly revenue per 100 buyers$918 +14%
Simulated from the conjoint: each respondent picks the plan, with or without Copilot Pro, that they value most at these prices, or none of them ("No plan bought" means none of these options, not a measured switch to a free or rival suite). Modelled separately for people who already pay for an AI assistant and people who don't, half of each. Copilot Pro buyers are counted under their plan, and its $20 is shown as add-on revenue; Copilot Pro bought on its own, without Microsoft 365, is not modelled. Prices per subscription per month, billed monthly; revenue is per 100 buyers in the survey. The result depends on that mix: with 10% already paying for AI, revenue changes +10% ($740 to $812); with 50%, +14%; with 90%, +18%. If the Classic plans are also on offer at the old prices, the change is +9% ($802 to $877) at 50%, with 55% of buyers choosing Classic. Early read: synthetic respondents, n = 300.

Result. Price drove 93% of the choices, plan size 6%, and the AI level only 2% overall, because the two groups pulled in opposite directions: people who already pay for AI preferred plans with AI, and the others preferred plans without it. Before the change, only 2.5% took Copilot Pro on top. After it, revenue per 100 buyers rises from $802 to $918 a month, +14%. But the share buying none of these plans goes from 4% to 12%, all of it among people without AI: 23% of them buy nothing at the new prices. Family falls from 26% to 12% of buyers.

For the decision: the $3 rise earns more than it loses, at any mix of AI and non-AI buyers (+10% to +18%). The cost is a large share of non-AI buyers walking away. Offering Classic at the old price keeps almost all of them, at a smaller gain (+9%), with most of them choosing it.

3

What is the acceptable price range for Personal with AI credits included?

Van Westendorp · all 300, then each group
0%25%50%75%100% $5$10$15$20 Personal at $9.99 Acceptable range $8–$12 · optimal ≈ $10 Too cheap A bargain Getting expensive Too expensive $5$10$15$20Already pay for AI$9$15Don't pay for AI$6$10○ optimal price
Top: share of all 300 who call each price too cheap or a bargain (falling), and getting expensive or too expensive (rising). The acceptable range runs from where "too cheap" crosses "getting expensive" to where "a bargain" crosses "too expensive"; the optimal price is where "too cheap" crosses "too expensive". Below: the same range for each group, n = 150 each. One-person plan, per month. Synthetic respondents.

Result. Across all 300, the acceptable range was $8 to $12, with the optimal price at about $10. So $9.99 sits right at the optimum. The groups differ: for people already paying for AI the range is $9 to $15, so $9.99 is near the bottom. For people without AI it is $6 to $10, with the optimum at about $8: $9.99 is the very top, and 67% of them already called a price of $9.99 or less "getting expensive".

For the decision: $9.99 is a fair price for the average buyer, cheap for AI users and the ceiling for everyone else.

4

At what price does Personal with AI credits bring in the most revenue?

Gabor-Granger · each group · prices $6.99, $8.99, $9.99, $11.99, $14.99
Share who would buy 0% 25% 50% 75% 100% Monthly revenue per 100 buyers $0 $250 $500 $750 $1,000 $6.99$8.99$9.99$11.99$14.99 Already pay for AI Don't pay for AI All 300 ◯ peak
Top: share who would buy Personal with AI credits at each price (demand curve). Bottom: price × that share, as monthly revenue per 100 buyers; the circle marks each group's peak. Per month. Synthetic respondents, n = 150 per group.

Result. Across all 300, revenue per 100 buyers is about the same at $6.99 ($671, 96% buying) and at $9.99 ($666, 67% buying). People already paying for AI peak at $9.99 ($966), with almost all still buying. People without AI peak at $6.99; at $9.99 only 37% would buy, and their revenue is 43% lower.

For the decision: asked about Personal alone, $9.99 makes about the same money as $6.99 from about a third fewer buyers. The buyers it loses are the ones a Classic plan at the old price could keep.

What the four add up to

The $3 rise pays, but it leans on the Classic plans to keep the people who don't want AI.

  • Value: people buy the subscription for the apps; AI matters only to those who already use it (question 1). Including AI credits for $3 more raises revenue by 14% in the simulation, but 23% of people without AI buy nothing (question 2).
  • Price: $9.99 is at the optimum of the overall range, and is where AI users' revenue peaks (questions 3 and 4).
  • Risk: for people without AI, $9.99 is the top of their range and only 37% would buy Personal at that price (questions 3 and 4). With Classic on offer, most of them take it, and the gain shrinks to +9% (question 2).

What this study can't tell you: whether people would find or ask for Classic, which Microsoft offered for a limited time, or what they would do instead. "No plan bought" is not a measured move to a free suite or a rival, and Copilot Pro bought without Microsoft 365 is not modelled. That is the next thing to test, with real respondents. Microsoft hasn't published what happened (see Result).

The full report · the PDF Albi produced for this survey: every question's chart, synthetic results labelled on every page PDF · 8 pages · download

How each method works, and how many respondents you need: Van Westendorp · Gabor-Granger · MaxDiff · Conjoint · Sample size

Find out what your buyers will pay, before you change the price

Run this survey yourselfStart from this Microsoft 365 survey and change it for your plans. $199 a survey, or $499 with conjoint, plus respondents: real people via Prolific, or synthetic.Use this survey →
Have Albi run itWe design, field and analyse a study on your own pricing question. Fixed price, invoiced.Talk to us about a study →

Microsoft 365 and Copilot are trademarks of Microsoft Corporation. Albi has no affiliation with Microsoft, and this teardown is not endorsed by it.

Sources

  1. Microsoft 365 blog, "Copilot is now included in Microsoft 365 Personal and Family", January 16, 2025: prices up $3 a month in the US, the first rise since launch; a monthly allotment of AI credits; Classic plans for a limited time; the stated reason. microsoft.com
  2. Computerworld, "Copilot AI comes to Microsoft 365 plans: Everything you need to know", January 22, 2025: annual prices $70 to $100 (Personal) and $100 to $130 (Family); about 60 AI credits a month; Copilot Pro at $20 a month. computerworld.com
  3. Microsoft Support, "Frequently asked questions about Copilot in Microsoft 365 subscriptions": AI features in Family are only available to the subscription owner. support.microsoft.com
  4. PCWorld, "Microsoft launches 365 Premium for consumers, retires Copilot Pro", October 1, 2025: Microsoft 365 Premium at $19.99 a month; standalone Copilot Pro discontinued. pcworld.com

We'd rather do this on your pricing than Microsoft 365's.

Same method, your own data behind it, and a proposal you approve or don't.