LoginSign up for free

Pricing AI features and agents by the work they do

The Pricing work group on billable events, credits, work packages and letting users try an agent before they pay.

Updated October 13, 2026

Bill for a unit customers can understand and verify, check it against the cost of serving heavy users, and let people finish a real task before they pay.

  • Define the billable event. Kyle Poyar cites Intercom's Fin at $0.99 per resolution, where the definition of a resolution matters as much as the rate. Dan Hockenmaier would define an outcome customers can verify, then test adoption at different prices alongside margin after delivery costs. Where outcomes are hard to verify, Alvaro Morales says tasks and workflows can be practical billing proxies.
  • Test how people buy. At Lovable, Elena Verna's team charges for building with credits and has no per-user fee in self-serve. It tested one-off top-ups at 20% and 40% premiums: 20% raised engagement and bookings, 40% had too few buyers, and paid retention improved 7% after the change.
  • Give buyers a budget. Kyle would sell an agent to a department head as a monthly package of work, such as the accounts an AI SDR researches, so the buyer can compare it with hiring. Alvaro points to Notion's then-$20 AI add-on, a simple price that let people try the feature, and tracks cost per customer behind it.
  • Let users finish a task first. Lauryn Motamedi would give a limited trial allowance so a customer can complete a real task, drawing on Airtable's reverse trial. Elena says to check first that users can see what the paid agent does and what it costs, so low conversion isn't blamed on the price.

Group chat with Kyle, Elena, Alvaro, Dan, Lauryn — the Pricing work group

Question:

How should you price AI features and agents?

Elena Verna
Elena VernaLeads growth at Lovable. Former SVP Growth at SurveyMonkey
I’d start with the unit of value customers can understand, then check whether the bill follows how they actually use the product. At Lovable, we charge for building with credits and let people collaborate without a per-user charge in self-serve. More collaborators can mean more building, so a seat fee would slow the behavior we want.LinkedIn
Agents make the unit harder to define. A single prompt can trigger a multi-step plan, so Agent Mode actions cost different amounts of credits depending on complexity. We told users the pricing would change and that we’d keep iterating.LinkedIn
Then test how people buy those credits. Lovable’s building was irregular, and subscription-only billing held back engagement. We tested one-off top-ups at the subscription credit price, then at 20% and 40% premiums. The 20% premium increased engagement and bookings; 40% had too few buyers. Paid retention improved 7% after the change. The practical test is engagement, bookings, and retention together, because a clean ARR line can hide a bad buying experience.elenaverna
Dan Hockenmaier
Dan HockenmaierPartner at Reforge. Former Chief Strategy Officer at Faire
The price needs to fund the growth engine. For an agent that completes a measurable job, model what the customer pays per completed job against the cost of serving heavy users. A base fee can give customers a predictable budget while covering fixed delivery costs.
In my pricing essay, I argued that the growth-optimizing price depends on where an extra dollar does more work: invested in sales, product, and support, or left with the customer to reduce friction and encourage use. Enterprise products that need expensive sales and delivery can justify more margin; products whose usage drives distribution benefit more from low friction.Medium
AI puts pressure on seat counts, and I’ve argued that seat pricing is often a sign the product’s ROI is hard to measure. For agents, define a billable outcome customers can verify, then test adoption at different prices alongside the margin after delivery costs.danhock
Kyle Poyar
Kyle PoyarFounder of Growth Unhinged. Former Operating Partner at OpenView
Intercom priced Fin at $0.99 per resolution. That works when customers and vendors can agree a support question was resolved; the definition of a billable resolution matters as much as the rate.Substack A generated video or completed research task may be easier to verify as delivered work, while downstream business success can depend on factors outside the agent’s control.
For an agent, I’d make the pricing decision in this order:
  1. Define the billable event and test whether customers agree it happened. A “successful resolution” needs a clear rule for handoffs, repeat contacts, and disputes.
  2. Estimate customer value per event and your cost per event, including model calls and human review. Set a price that leaves room for both customer ROI and your gross margin.
  3. Give buyers a way to budget. A base subscription with included usage and a clear overage rate can make variable consumption easier to buy. For an agent that resembles a specific role, package a predictable volume of work against the cost of that role.
In our review of 40 leading AI apps, Palle Broe and I found seven in ten used subscriptions, and most charged per user. That fits products working alongside a person.Substack I’d revisit the metric when the agent starts completing work independently.
Elena Verna
Elena VernaLeads growth at Lovable. Former SVP Growth at SurveyMonkey
One step before testing the price: check whether users can see what the paid agent does. In self-serve, I’ve seen teams put a paid feature in a higher plan and make it invisible to everyone else. Then low conversion gets blamed on price when users never had a chance to consider the feature. I recommend reviewing the paid experience in every plan state, including free, and measuring awareness before changing the price.Spotify+1
For an agent, show the task it can complete, the credit cost before someone starts, and what happens when credits run out. Then separate three questions in the data: did people discover it, did they understand the charge, and did the price stop them from using it? That sequence gives a pricing experiment a fair shot.
Lauryn Motamedi
Lauryn MotamediFormer VP Growth and Product at Notion and Head of Growth at Airtable
Give users a chance to finish a real task before deciding where the paid boundary sits. At Airtable, I argued for a reverse trial: let someone try premium capabilities while they learn the product, then give them a lasting free path after the trial. The point was to help them experience value before asking for a card or a sales call. I also cautioned against steering new users toward a premium feature simply because its users convert more often; onboarding should fit the user's job.YouTube
For an AI feature or agent, I would apply that logic with a limited trial allowance that lets a customer complete a real task. Then I would price the recurring use against the value of that task, with clear usage limits where running it costs more. Track task completion, repeat use, paid conversion, and serving cost by customer segment before settling on the allowance or price.
Alvaro Morales
Alvaro MoralesCo-founder and CEO of Orb. Former Engineering Manager at Asana
Price an early AI feature for the behavior you need to observe. Adoption and customer value may matter more than maximizing revenue per request. I’ve pointed to Notion’s then-$20 AI add-on as an example of a simple fixed price that let people try the feature while the company learned how it affected customer lifetime value. The trade-off is that heavy users can cost more to serve, so measure usage and cost per customer even if the customer sees a flat price.Spotify
Kyle’s resolution example shows why the event definition matters. For agents whose outcomes remain hard to verify, tasks and workflows can be practical billing proxies. In enterprise deals, a prepaid pool of credits can give customers a spending commitment while allowing usage to vary.Apple Podcasts
Then make the usage and expected charge visible while the customer is using the agent. A customer should be able to see the cost before the month-end invoice. Keep the billing system flexible enough to change the metric or combine seats and consumption as you learn which customers get value.Apple Podcasts+1
Kyle Poyar
Kyle PoyarFounder of Growth Unhinged. Former Operating Partner at OpenView
For an agent sold to a department head, the price page can translate variable tasks into a predictable work package. I’ve seen buyers struggle to forecast a pile of AI credits; a monthly package tied to the work of a specific role gives them a budget they can compare with hiring.Substack
For example, an AI SDR package could state how many accounts it researches, messages it sends, and replies it handles each month. The underlying bill can still track tasks or usage. In my conversation with 11x founder Hasan Sukkar, he described task-based pricing, while the buying pitch compared an AI SDR’s output with a human SDR at roughly one-fifth the cost. That comparison only holds if the agent delivers useful output at a comparable quality, so I’d show the included task volume and overage rate beside the headline price.Substack

Ask a follow-up

+2

Pricing

Ask Pricing anything...
+2

105

Pricing

3M+ insights

Pricing & monetization experts from Faire, Amplitude, Dropbox & 1 more

by SuperMe

In This Conversation

  • Kyle Poyar
    Kyle Poyar

    Founder of Growth Unhinged. Former Operating Partner at OpenView

  • Elena Verna
    Elena Verna

    Leads growth at Lovable. Former SVP Growth at SurveyMonkey

  • Alvaro Morales
    Alvaro Morales

    Co-founder and CEO of Orb. Former Engineering Manager at Asana

  • Dan Hockenmaier
    Dan Hockenmaier

    Partner at Reforge. Former Chief Strategy Officer at Faire

  • Lauryn Motamedi
    Lauryn Motamedi

    Former VP Growth and Product at Notion and Head of Growth at Airtable

See all

Frequently Asked Questions

  • How should a consultant or agency price its services?

    By the shape of the work: a fixed fee for a defined deliverable, a retainer for ongoing access, prepaid units for variable requests. Add a performance fee only once the metric, baseline and client duties are agreed.

  • How does a young startup or community choose a business model?

    Find the behavior that creates value and charge where it does the least damage to it. Test a small paid offer with real buyers, then check they keep paying after the first purchase.

  • Should a transaction fee be flat or a percentage?

    Flat when the work and value per transaction are about the same, a percentage when value or risk grows with order size, and volume tiers when each extra transaction is worth less.

  • How do you present a pricing change to executives?

    Open with the decision, show each account's price change next to its health, and ask for a staged test with a return date and stop rules agreed in advance.

  • How many pricing tiers should a SaaS product have?

    Most end up with three to five plans, and Good-Better-Best is the usual default. Choose the value metric first, then step each tier up with the customer's sophistication.

  • How do you break down a competitor's pricing?

    List each plan's model, value metric and upgrade triggers, use the product yourself, then work out where the revenue sits and whether a high price pays for sales or is margin.

  • How do you move existing customers to a new pricing model?

    Simulate each account's first bill, test the model on new buyers, then move existing customers in cohorts with notice, a fallback plan and a migration date that respects renewals.

  • How long should a free trial be?

    Long enough for retained customers to reach the paid value. For a simple self-serve product, start with 14 days as a test, then compare 14 and 30 days on revenue and retention at the same cohort age.

  • What makes people pay at a paywall?

    A wall that stops them partway through a job they care about, a limit on the unit they value, an offer that fits their usage, and copy and checkout that don't leak buyers.

All pricing questions  →

Your ideas, super powered.

Put what you know to work for others, and think through your hardest questions with people who've been there.

Topics

  • Startup growth
  • Job search
  • Starting a startup
  • Career growth
  • Product management
  • Building with AI
  • Pricing
  • Leadership and management
  • Paid acquisition
  • B2B sales

Popular Questions

  • Choosing the first growth channel for an early startup
  • Preparing for PM metrics interviews by deriving the metric out loud
  • Reading the early signs of product-market fit
  • What to check before you accept your next role
  • The choices a good product strategy has to make
  • A first LLM eval built on a graded sheet, a bug set and a launch bar
  • Retainers, project fees or upside: pricing consulting work
  • Where a first-time manager should put their attention
  • Scaling paid social past a plateau
  • Setting up a startup CRM around what buyers do

Work Groups

  • AI Product
  • Founders
  • Executive Coaching
  • Product Leaders
  • Product Careers
  • Growth
  • Pricing
  • Performance Marketing
  • Product Interviews
  • B2B Sales

© 2026 Super Me Network, Inc.

Privacy Policy

Terms of Service

contact@superme.ai