Asana AI Studio Pricing: Credits per Plan and What Runs Out First

Asana AI Studio is included on paid plans: Starter provides 50,000 credits, Advanced 75,000, and Enterprise or Enterprise+ 200,000 per billing account monthly. Starter starts at $10.99 per user monthly on annual billing. AI Teammates and Dash use a separate request pool: five per user, capped at 50 on Starter and Advanced. Your workflow determines which runs out first.

The two meters are easy to confuse. A smart rule that reads a task and writes a short field update spends AI Studio credits. An AI Teammate completing work, or a qualifying Dash action, spends an AI request. Buying requests does not replenish Studio credits. For a small team, a busy Teammate can hit its pooled request limit even while most Studio credits remain.

Follow me on Instagram@sagnikteaches

The price and included allowance by Asana plan

Asana's public pricing page lists Starter at $10.99 per user per month when billed annually, or $13.49 on monthly billing. Advanced is $24.99 per user per month annually, or $30.49 monthly. Enterprise and Enterprise+ require a sales quote. These are plan prices; a separate AI Studio Plus purchase or paid AI requests increases the bill. Check the checkout screen before committing because billing cadence and promotional terms can change.

Connect on LinkedInSagnik Bhattacharya

Notice the different units. Five Starter seats produce 25 included requests for the entire account, but they do not produce 250,000 Studio credits. The Starter account still receives 50,000 credits. Conversely, 20 Starter seats reach the 50-request cap; multiplying 20 by five would be wrong. Asana's product terms say requests are aggregated and can be consumed by any user on the account. A heavy user may consume more than five personally while the account pool remains available.

Subscribe on YouTube@sagnikteaches

Advanced's extra 25,000 Studio credits should not be mistaken for a cheaper way to buy a large AI capacity increase. Five Advanced seats cost $124.95 per month on annual billing, compared with $54.95 for five Starter seats: an illustrative $70 monthly difference before any add-ons. Advanced also adds portfolios, approvals, form branching and resource features. If those are useful, the higher base plan may make sense. If only Studio credits are constrained, price the relevant add-on against that $70 difference.

Estimate a rule from input, output and repeat volume

Asana's credit guidance says that input size, output size, selected model and execution frequency determine consumption. It gives a planning formula: input tokens times the model's input multiplier, plus output tokens times its output multiplier, divided by 100. Roughly 100 tokens make one credit before those multipliers. A token is a chunk of text the model processes, so a long task history or attachment can cost much more than a short title. Treat this as an estimate, then inspect the rule's real execution history.

For example, Asana's published reference lists a lighter model with a 0.5 input multiplier and 5 output multiplier. A rule reading 1,200 input tokens and returning 120 output tokens would use approximately (1,200 × 0.5 + 120 × 5) / 100 = 12 credits. If it fires 1,000 times in a month, that is about 12,000 credits. The same input may cost several times more with a heavier model. Model availability and recommended defaults can change, so confirm the selected model in the rule editor rather than budgeting from a remembered setting.

Ask the person building the rule to show the exact task fields, attachment types, linked projects and expected output. A short classification such as “urgent”, “standard” or “needs review” can be far cheaper than a paragraph of explanation that nobody reads. In an illustrative security installer, a rule that routes 600 monthly service enquiries needs the fault type and requested date; pulling a long discussion thread and every site photograph into the prompt adds cost without improving the routing decision.

Here is compact rule guidance for that security installer's routing rule, followed by an illustrative test result and the fix it needed.

Read only the task title and the fields Fault type, Requested date and Site access.
Set the field Queue to one of: Urgent call-out, Standard service, Needs review.
Write one sentence in the Handover field naming the fault and the requested date.
If Fault type is empty, choose Needs review and say which field is missing.
Do not use attachments, comments or linked projects.

Illustrative output on a test task: Queue set to "Urgent call-out"; Handover reads "Alarm panel showing a tamper fault; customer confirmed for Thursday morning." The queue was right, but the rule had turned a requested day into a confirmed booking. The fix was one extra line telling it to write "requested" before any date, because the office, not the rule, confirms appointments. The test also showed that a one-sentence output keeps the output side of the formula small. Asana's credit guidance names @mentioned projects and the rule setting that includes attachments and links added to tasks as the biggest input levers, so leave both out unless the decision genuinely needs them.

The practical measuring method is to run ten representative tasks through the workflow. Include short, ordinary and unusually long cases. Record the credits each consumed in rule execution history, then calculate an average and a high case. Multiply both by realistic monthly volume. Do not rely only on a perfect demo task: the long exception is often where attachments, repeated retries or verbose output stretch the estimate. For a related way to limit intake before AI sees it, see the status update workflow, which starts from structured task data.

A five-person team's month, calculated twice

Consider an illustrative five-person signage company on Starter. It has 800 incoming production requests each month. A Studio rule classifies each request using a short brief and produces a two-line handover; pilot executions average 18 credits. A second Studio rule runs only on 120 unusual requests and averages 90 credits because it reads a longer brief and drafts a clarification. The team also uses AI Teammates to complete 32 separate follow-up jobs. These numbers are invented for arithmetic, not Asana's guaranteed price per task.

In this run, the request pool runs out first by seven requests, while 24,800 Studio credits remain. The team could buy seven on-demand AI requests at $0.60 each, an illustrative $4.20, if on-demand billing is enabled and its spend cap allows that use. Prepaid requests are listed at $0.50 each but require an upfront purchase. If it buys no capacity and does not enable on-demand billing, Asana's product terms say features that need AI requests stop being available until more are bought or on-demand charges are authorised; the included pool refreshes monthly, and the Studio credit balance is a separate meter.

Now change just one assumption. If the production rule accidentally runs twice on every intake edit instead of once at submission, its monthly 14,400 credits become 28,800. The total rises to 39,600, still below Starter's 50,000 but with far less room for long briefs. If the exception rule also fires twice, total use reaches 50,400. The same five seats and 32 requests now face both meters. That is why the trigger event deserves as much scrutiny as the prompt wording.

For an interior designer with only 90 substantial project briefs, a heavier rule might be affordable because it runs rarely. If each run costs an illustrative 250 credits, the month's use is 22,500. In a security installer processing 2,000 short service notes, even a compact 15-credit rule uses 30,000. A solar installer whose rule reads 40-page attachments should pilot those large documents separately; an average based on short enquiries would hide the costly case. These are different workflows, so there is no honest universal “credits per task” figure.

Which limit should you watch on Monday morning?

Start with the jobs people perform, rather than asking which meter is bigger numerically. List each AI action in the team for a week and label it Studio rule, Teammate, Dash or ordinary non-AI automation. If the activity is ambiguous, use Asana's admin usage dashboard and a test execution to see which meter changes. Asana's pricing FAQ says administrative tasks in Dash are included at no extra cost and that Dash only counts a request when it completes complex work, while each successful unit of Teammate work counts. The live record is more reliable than assigning every AI-looking button to the request pool.

On Starter and Advanced, ten users create 50 included requests, at the account cap. If each of those users asks a Teammate to complete two jobs a week, the team might need about 80 in a four-week working month. That is an illustrative 30 above the included pool, even if Studio is barely touched. Upgrading from Starter to Advanced does not raise this 50-request ceiling. Compare the real extra request cost with the broader value of Advanced before changing the base plan.

Credits are more likely to run out first when a high-volume rule repeatedly handles long inputs, generates long answers or uses a premium model. One illustrative 300-credit rule run 200 times spends 60,000 credits. The same team might use only ten Teammate requests. An owner who sees “only ten AI jobs” could therefore misread the bill: those 200 automated rule runs are a separate meter. The small-team Asana cost comparison helps place AI usage alongside seats rather than treating AI as the whole subscription.

Set a weekly review that records opening balance, credits spent by the highest-volume rules, included requests remaining and any on-demand charges. Weekly reviews catch trigger mistakes in time to edit them. Month-end reviews explain a surprise but cannot undo a month of repeated executions. If consumption jumps, compare the rule's execution count first, then inspect average cost per run; these distinguish a volume problem from a prompt or model problem.

What to change before buying Studio Plus

First shorten the input. If a rule only assigns an owner, feed it the service category, deadline and location code, not an entire contract or project thread. In an illustrative signage company, the original prompt read a 2,400-token customer history and returned a 300-token explanation. After changing it to read 400 tokens of intake fields and return a one-word queue label plus one sentence, a pilot can show whether quality stayed acceptable. Measure the actual credit reduction; do not assume the percentage from text length alone because output and model multipliers differ.

Second, reduce the number of times the rule fires. A trigger on “task updated” can re-run when a colleague corrects a spelling, uploads an attachment or changes an unrelated field. A trigger tied to a specific section or field transition may run once per stage. Keep a human review queue for exceptions. For the ordinary assignments, a plain Asana rule may be enough; five starter automation rules offers a useful comparison before AI is added.

Third, choose a lighter model for classification and routing. Keep a more capable model for the few cases that need reasoning or nuanced language. Asana's credit guidance says to start with the lightest model that produces acceptable results, and because the model is chosen rule by rule, you can mix them. Run the same representative tasks through both choices and score the outputs: correct category, correct owner, no invented promise, readable handover. If the cheaper model fails on uncommon safety cases, route those cases to a person rather than paying for a heavy model on every ordinary request.

Fourth, reduce unnecessary output. One solar installer might ask a rule to produce a 500-word summary for every quote enquiry, yet the team only needs roof type, requested system size and missing documents. A compact field update may make work clearer as well as cheaper. Keep full prose for the small fraction that needs a customer-facing draft. If customer copy is generated, make a person check any numbers and commitments before it is sent.

Only after those changes should you price AI Studio Plus. Asana says Plus credits stack with the Basic allowance and are bought as 100,000-credit packages from the Billing tab of the admin console, up to 1 million credits in one purchase; its help material says Plus credits expire monthly with no rollover. A Starter billing account with one 100,000-credit package would have 150,000 for that month. The amount due depends on the current checkout and billing cadence; the public help page explains the purchase mechanism but does not give a universal transaction price. Ask the billing owner to capture the checkout total and the renewal date before approving it.

Paid requests have their own cost controls

Asana lists prepaid AI requests at $0.50 each and on-demand requests at $0.60 each. Prepaid packs and on-demand use have different timing and controls. On-demand billing can be given a monthly spend cap; Asana's live dashboard shows used and remaining requests, and its pricing FAQ describes alerts at 65%, 85% and 100% of the allowance. For a team whose use is uncertain, a modest cap can prevent a surprise bill while still allowing a small overrun. Timing matters for packs too: the pricing FAQ says packs bought on a monthly plan expire at the end of the month, while packs on an annual plan can be used across the contract term.

An illustrative five-person office expects 25 included requests and forecasts ten extra. At the listed on-demand rate, that is $6 for those ten; prepaying ten at the listed rate would be $5, subject to the available pack purchase terms. The dollar difference is small, so the control question matters more: does a stopped Teammate block a customer handover, or can the work wait for the next refresh? Do not buy a large prepaid quantity merely because the unit price looks lower.

Watch who can start paid work. If a colleague repeatedly asks a Teammate for a slightly different draft, the successful completions may count separately. Define the task, expected output and approval point before using the pool. Asana's pricing FAQ says failed, timed-out or cancelled tasks do not count, but the admin usage record should settle any billing dispute. For team processes that already work in ClickUp, compare the three project tools' AI workflows without assuming their meters are interchangeable.

A simple purchase test for the next billing month

Use a one-page record with five figures: paid seats, included Studio credits, measured credits per rule run, monthly trigger count and successful AI request count. Put the billing reset date beside it. For each rule, show an ordinary and a high-case cost. Add a short note about what human check remains. This turns a vague “we need more AI” request into a purchase decision the owner can review.

For the illustrative signage company, the record would read: five Starter seats; 50,000 included credits; classification 18 credits × 800, exception draft 90 × 120; 25,200 expected credits; 32 AI requests against 25 included. The immediate fix is to decide how to cover seven extra requests, then monitor the rule trigger. Upgrading to Advanced just for the request overrun would be poor arithmetic. If the rule usage grows beyond 50,000 after optimisation, quote Studio Plus separately.

Run the same calculation with a high-case month. Suppose 1,200 intake requests and 180 exceptions arrive, with unchanged credit averages. Studio use becomes 21,600 plus 16,200, or 37,800. If each exception grows from 90 to 180 credits because more attachments are read, total becomes 54,000. That crosses Starter's included allowance without changing seats or requests. The high case tells you when to shorten input, route attachment-heavy cases to people, or purchase capacity before the busiest period.

Finally, check the admin console against your spreadsheet each week. If measured use is higher, inspect the exact rule history and update the assumptions; if it is lower, keep the smaller plan until the work proves otherwise. The pricing page tells you the allowances, while your own execution log tells you which meter will run out first in your business.

Questions before buying more Asana AI capacity

Do unused AI Studio Basic credits roll over?

Treat the included Basic allocation as a monthly allowance, not a balance you can bank: Asana's product terms say included credits refresh monthly and unused ones cannot be carried into a new subscription term. Its help material describes Plus credits as expiring monthly with no rollover. Check the billing account's reset date and the current admin console before adding a paid package, especially if you are testing late in a billing period.

Can one heavy user consume the whole AI request pool?

Yes. Asana calculates included requests from paid users, then pools them at organisation level. There is no individual five-request cap. One person using AI Teammates or qualifying Dash work can consume requests that other colleagues expected to use. Review the live usage dashboard and decide who should have request access.

Does upgrading from Starter to Advanced solve every AI limit?

No. It increases included AI Studio credits from 50K to 75K per billing account, but a small team's pooled AI requests remain five per user, subject to a 50-request account cap. Upgrade for the broader project features too; compare a Studio Plus package if credits alone are the problem.

Further reads

• ClickUp Brain Pricing for a Small Team: Brain AI vs Everything AI — Compare another project tool's AI pricing model.
• Asana to ClickUp Migration: Import Steps and What Breaks — Assess the migration work if pricing prompts a switch.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: Asana pricing page and FAQ (asana.com/pricing); Asana Product-Specific Terms on AI Requests and AI Studio credits; Asana help guidance on AI Studio credit optimisation and Studio Plus; Asana community announcement of AI Studio Plus direct purchase (checked 5 November 2026).