10 AI pricing moves you can steal this week
Plus: Updates from Clay, Synthesia, Hubspot, Vercel, and Miro.
Welcome back to Good Better Best.
Each week, we break down real pricing, packaging, and product moves from SaaS leaders and share the ideas worth stealing.
We’ve seen a flurry of AI-related pricing changes over the past few weeks, so today curated some interesting examples for everyone out there navigating the options. For further research, check out PricingSaaS Pulse, which is currently in free beta.
Warning: this post went too long for email, so click ‘View in Browser’ to read the whole thing. Let’s get to it.
PS. Next week, we’re hosting an Office Hours session with Manny Medina, CEO of Paid on all things AI credit models. Learn more and register here.
PPS. My friend (and former boss 🫡) Jordan DiPietro put together an awesome guide to decoy pricing based on an experiment he ran at Motley Fool that generated $8M in annual revenue. Pricing psychology still works — especially for PLG/consumer products. Grab it here.
🔌 PricingSaaS Partners power the next era of SaaS pricing
This Week in Pricing, Packaging, and Product
We’ve ramped from 500 to 2,000 companies tracked, and predictably, change volume followed suit. Last week we observed 120+ changes. The highlights:
Clay updated its pricing model [Link]
Synthesia hiked the Starter plan 53% to $29/month [Link]
Miro shifted Enterprise AI credits from 100/member to custom amounts [Link]
Vercel launched Queues Beta with 1M free ops, $0.60 per 1M extra [Link]
SimilarWeb launched AI Studio for natural language market insights [Link]
Softr redenominated AI credits: 100 old credits equal 1 new credit [Link]
Hootsuite removed Approval Workflows and Permissions from Advanced [Link]
Dialpad replaced Connect with AI Agent conversation pricing [Link]
HubSpot added User Management to Free Tools [Link]
Supermetrics added Management API to Enterprise plan [Link]
Check out more updates on PricingSaaS →
PricingSaaS Pulse Intelligence
Here’s what was top of mind in Pulse this week:
🔥 Hot Companies
Klaviyo — 19 searches
Zendesk — 16 searches
Notion — 15 searches
Intercom — 12 searches
Stripe — 12 searches
🚨 Hot Topics
Agent vs. Assistant pricing — OBP vs. UBP, complexity pitfalls
AI monetization — credit-based pricing, bundling strategies
Free beta → paid transition — customer backlash, change management
Enterprise packaging — paid POC, proof of concept pricing, land & expand
Commercial debt — price anchoring, free features, grandfathering
10 AI pricing moves you can steal this week
For the past year, everyone’s been asking the same question: how do you price AI?
I wish there was a silver bullet. The honest truth is it depends. It’s hard to draw hard trends across the wider SaaS landscape. We’re talking about thousands of companies that serve hundreds of different buyers at varying levels of AI proficiency.
In the last few weeks, we’ve seen a wide range of pricing, packaging, and product changes. Below we’ve curated a punch-list of tactics that are worth watching.
1️⃣ Monday.com’s Enterprise AI Bundle
Enterprise plans are usually pretty boring. But Monday.com just did something interesting to change their. They've bundled all AI features - Sidekick, AI credits, and Vibe - into an "Enterprise AI bundle" that only lives on the top-tier plan.
This could have an interesting impact on the upgrade conversation. The old enterprise motion was driven top-down. IT needed better permissions. Security needed compliance controls. The decision happened in a room full of people who were not using the product. This enables a bottom-up approach. [Link]
2️⃣ Notion’s Custom Agents Add-On
A couple weeks back, Notion launched Custom Agents as a paid add-on. It’s available only to Business and Enterprise customers, who can trial for free, then pay on a credit-based model ($10 per 1,000 credits).
The “s” in agents is the important here. Given Notion’s broad functionality, there’s a high likelihood heavy users will create multiple agents that do various jobs, making the add-on a strategy a no-brainer. [Link]
3️⃣ Miro’s Custom Enterprise AI Credits
Miro has a tiered AI credit system, and this week they changed Enterprise from 100/member to “custom amounts.” Removing the cap gives Enterprise prospects flexibility, and forces a conversation to determine the right allotment. [Link]
4️⃣ Pendo’s Flexible AI Add-On
Pendo restructured their entire pricing page — putting AI front and center. They offer baseline AI functionality across all plans, while allowing customers to add more AI modules as needed. This flexible approach aligns with their pricing philosophy which is anchored in All-In Access (no feature gates within a purchased module) and Customization. [Link]
5️⃣ GitLab’s Agent Integration
GitLab’s AI pricing move isn’t about their own models — it’s about becoming the platform layer for third-party AI agents. By adding Claude Code and Codex to their Duo Agent Platform, GitLab is positioning itself as the central hub for agentic development workflows. [Link]
6️⃣ SimilarWeb’s AI Studio Module
If you’re investigating the modular approach, SimilarWeb is another interesting example. They recently repackaged their AI features and launched a new module called AI Studio. This allows them to offer an AI-native layer on top of their existing dashboard-driven analytics, and capture additional revenue. Particularly useful given their sales-led approach for Business and Enterprise accounts. [Link]
7️⃣ Dialpad’s Conversation-Based Agent
Dialpad pulled from the Intercom playbook and introduced a new AI Agent tab to their pricing page, differentiating it from existing plans with ‘per conversation’ pricing. This gives them a seat at the table for OBP-based Customer Success products, and is a smart retention play to engage customers who are exploring OBP. [Link]
8️⃣ Rocketlane’s Enterprise Agent
Rocketlane introduced two agents, differentiated by credit volume, and the actual jobs the agent can do. This feels like a peek into the future, where more and more plans will be differentiated by agent capabilities, and “harder jobs” garner a higher price tag. [Link]
9️⃣ Softr’s AI Credit Reduction
Softr cut their free plan AI credits by 99% — from 500 to just 5. Expect more of these pullbacks as companies realize free AI credits are a line item, not a rounding error. The key is finding the limit where free users actually see enough value to upgrade. This is also why I’m bullish on MCP as a freemium strategy (it takes LLM costs off your books). [Link]
CORRECTION: Softr actually redenominated AI credits. 100 old credits equal 1 new credit. Customers get the same level of usage. More information here.
🔟 Hubspot’s Free Trial for Customer Agent
Hubspot launched a 28-day unlimited free trial for its Customer Agent, giving Professional and Enterprise users the ability to fold it into their workflows.
Free trials remain underrated assuming you have the CSM team to support it. It’s a smart move to convert early customers and gather learnings to make the product better long-term. [Link]
Looking Ahead
If there’s one takeaway from these moves, it’s that the “how do we price AI?” question is being answered in real time — through bundles, credits, modules, agents, and good old-fashioned free trials.
There’s no bulletproof strategy, only iteration. If you’re working on this and could use a sounding board, shoot me a note at rob@pricingsaas.com.
Thanks for reading! When you’re ready, here’s how we can help:
Community: Answers from real pricing experts. Apply for membership →
Data: Access to news, research, and analysis. Sign up for free →
Advisory: Free pricing advice. Book a 30-minute call →














I believe Softr just revalued their credits. I received this via email:
3. Credit denomination update (no price change). We're also simplifying how AI credits are displayed.
- Previously, 100 credits represented 1 unit of usage.
- Going forward, 100 old credits will equal 1 new credit.
This is purely a denomination change to make usage easier to understand and manage. There is no change to pricing or the total value of your credits, and your purchasing power remains exactly the same. Only the way credits are counted and displayed has been simplified. If you have any questions, we're always happy to help.
It is interesting to compare the price per credit number across solutions. There is huge variation here. I am wondering if many people bother to calculate the price per credit, how it changes across tiers and with volume, and what it signals. I am see ranges of $5 per credit to fractions of a penny per credit. The number by itself does not mean a lot of course, as there is also a great deal of variation as to what a credit gets you. I plan to research this over the next few days and see what patterns if any I can find. Have you looked at this?