How to make PLG + SLG work in the AI era
Plus: Updates from Zoom, Anthropic, Algolia, Vidyard, and Triple Whale.
Welcome back to Good Better Best.
Each week, we break down real pricing and packaging moves from SaaS leaders and extract the ideas worth stealing.
This week we hosted an office hours session on one of the thorniest questions in SaaS: how to actually make PLG and SLG work together in the AI era. Below, I shared 7 key takeaways, plus the usual dose of pricing, packaging, and product changes.
Let’s get to it.
🔌 PricingSaaS Partners power the next era of SaaS pricing
This Week in Pricing, Packaging, and Product
Zoom raised the price of AI Companion.
Anthropic released Opus 4.6.
Jam reduced AI Generation limits.
Algolia added new generative features.
Vidyard reduced free video limits.
Triple Whale introduced AI Credits.
Check out more moves from Vercel, Frontegg, and Snyk on PricingSaaS →
This week, we hosted a live Office Hours session with Mark Walker and Tina Kung, CEO and CTO of Nue.io to answer one of the thorniest questions in SaaS right now:
How do you make PLG and SLG work together in the age of AI?
In the AI era, it often feels like we’re watching a giant game of musical chairs. Users are signing up for everything, and committing to nothing. The companies that can convert self-serve curiosity into contracted, committed revenue are the ones that will still be here five years from now.
Mark and Tina aren’t armchair theorists on this, they’ve spent decades in the trenches of subscription and revenue operations.
Mark is a serial entrepreneur who’s spent 20+ years building and scaling SaaS companies, with a particular obsession around CRM and ERP systems. He’s seen what happens when pricing and billing systems break under pressure.
Tina is the architect behind Nue’s technology. She’s built four generations of CPQ and billing systems across Oracle, Salesforce (Steelbrick), and Zuora, and led product and engineering for one of the largest CRMs in China.
Together, they’re building Nue to solve a problem they kept running into: the tools built for subscription businesses were never designed for the messy reality of PLG and SLG coexisting in the same company. Nue is now powering these motions at top players in SaaS and AI including Superhuman, Glean, OpenAI, and Chili Piper.
The Nue team recently released their Guide to Commit Burndown, which is quickly becoming the revenue model powering the AI era. I highly recommend bookmarking it for later.
The questions from our community didn’t disappoint. We had submissions from leaders at Monday.com, Zoom, and Typeform. Here’s what we learned 👇
Takeaway #1: The PLG/SLG binary is false.
Most companies will operate in both motions simultaneously, and the customer experience should flow seamlessly between them.
Both Mark and Tina argue that the framing of PLG vs. SLG as two separate worlds breaks in practice. Their advice? Design for fluid movement between motions. Mark went further, breaking down why the distinction doesn’t hold on either end.
The distinction between PLG and SLG was never real anyway. Very few companies reached massive scale on pure PLG. Eventually, either customers’ CFOs want control, or you start dealing with enterprise tiers that require security reviews and formal contracting. AI is accelerating this. Data concerns and auditing tools will force SLG conversations. On the other end, the PLG/SLG boundary breaks down post-sale as well. Enterprise customers, once contracted, generally don’t want to talk to a salesperson again for routine expansion. They want self-serve interfaces for adding volume or understanding consumption.
Takeaway #2: Flexibility is the meta-strategy.
In the AI era, product roadmaps, pricing models, and customer needs change too fast for rigid packaging. Build for adaptability.
Mark pointed to the leading AI companies as proof that flexibility in packaging is the defining strategic priority. Nobody knows what future pricing will look like, so the infrastructure must be designed for constant change.
Companies shouldn’t pretend they know what future pricing will look like. The AI leaders: OpenAI, Anthropic, Cursor, Harvey, Anaconda — all adopted flexible systems precisely because they don’t know what the pricing and packaging mix will be in the future. That humility is the strategy. Build packages and contracts structured to deliver increasing value without requiring re-contracting. The message to customers should be: we’re going to be dedicated to bringing more value to you every month, every quarter, and this contract is structured so you can grow as fast as possible without going through a re-contracting process.
Takeaway #3: Credits and committed spend are the bridge.
They preserve PLG simplicity while enabling SLG complexity, and let customers gradually adopt new services without re-contracting.
Credit-based models were a focus of the session, with good reason — PricingSaaS data showed 126% YoY growth in credit models in the past year. Tina gave historical context for credit models in SaaS, and how AI companies are giving it a new spin.
Snowflake’s journey is the canonical example. They innovated the credit burn-down model to create a common unit across compute, storage, and API usage. In the modern AI era, the evolution is toward committed spend — you commit an amount and pay as you go rather than prepaying a lump sum. This lets customers gradually adapt while giving sellers predictable ARR. The committed spend can continue from PLG subscriptions and gradually incorporate more services: subscriptions, AI tokens, professional services, FDE (Forward Deployed Engineers), with no big contract required upfront. Customers commit more or less depending on service needs, with pricing that appears simple on the surface.
Takeaway #4: Professional services are a growth lever, not a cost center.
Embed them in committed-spend contracts, make them self-serviceable, and think of them as usage products.
Mark and Tina share a contrarian perspective on professional services. Instead of a gated implementation requirement, they view professional services as a flexible, revenue-generating product.
Think of professional services as a usage product. It has a price, and can be structured in hour blocks. An emerging pattern is the “go-to-market engineering” function, where professional services is the sales team. They go in, configure and deliver the solution, and that installed solution then becomes the PLG expansion vector. You need professional services to get the product installed and change business processes before self-serve expansion can take hold. Major AI companies are embedding professional services inside committed-spend contracts. Customers don’t have to re-contract, they just raise their hand and get billed against the existing agreement.
Tina elaborated — highlighting why professional services are becoming a necessary complement to rapidly evolving products.
In the AI era, selling a solution is not a one-time event — it’s ongoing. Products update constantly, and managed services or FDE motions need to be self-serviceable. Customers should be able to add professional services hours on demand for new product rollouts or consultative support, without renegotiating. Ideally, they are bundled alongside subscriptions and consumption — creating a unified revenue model that includes one-time, milestone-based, and recurring billing — rather than being a separate gated engagement.
Takeaway #5: Don’t cancel-and-replace.
Whether transitioning a customer between motions or introducing new pricing, preserve continuity. Friction kills expansion.
Tina argues that preserving the customer’s existing subscription and experience is tablestakes to a frictionless transition between PLG and SLG.
PLG customers are accustomed to transparent, simple pricing. When moving to SLG, pricing gets more complex — multi-year contracts, discounts, professional services add-ons — but the challenge is maintaining that sense of transparency and simplicity even as complexity grows underneath. From the customer’s perspective, they shouldn’t feel like they’re canceling one subscription and replacing it with a completely different one. And because there may be a motion to move back, the back-and-forth should be very frictionless.
Mark went even further, highlighting the risks of a poor pricing model or billing process when converting PLG customers to higher plans.
Customer dissatisfaction with the pricing model or billing process — not the software itself — is one of the main drivers of churn. Offering different value-aligned pricing options, rather than forcing a switch, addresses this directly. There’s a key difference between the absolutist approach (stay on old pricing or switch entirely to new) versus the incentive approach (old model has limited features at higher cost, new model offers more value with better economics). The second approach creates a re-engagement opportunity rather than a disruptive migration.
Takeaway #6: Involve finance early.
The CFO’s perspective on revenue recognition, SKU management, and reporting should shape packaging decisions from the start.
Revenue recognition, reporting, and billing complexity all need to be factored in before new models ship. If not, Mark says, you’ll end up with a laundry list of SKUs that need to be normalized later in reporting.
Talk to your CFO. A big issue is whether these are actually different products or different pricing plans on the same product. You want the same product in different offerings, maybe with a completely different pricing model, but with revenue going to the same place. Otherwise, you’ll configure your financial system backwards and create more complexity.
Takeaway #7: System architecture matters.
Disconnected billing, CPQ, and self-service systems are the bottleneck. Unified platforms are required to execute these strategies.
The final takeaway is that none of the preceding strategies work without the right system architecture. Tina argues the PLG-to-SLG problem is fundamentally an infrastructure problem.
Revenue recognition naturally starts from how you price and assign revenue models for your products. It starts from products and pricing, flows through quoting, billing, and rev rec — these must be connected and unified. PLG and SLG have very different metrics (PLG tracks conversion and usage; SLG tracks ARR and ACV), and the analytics layer must unify these across channels. When PLG and SLG have different billing systems (e.g., Stripe for PLG, another for SLG), the entire ecosystem needs unified traceability and analytics. Without this, you lose clarity on what’s driving revenue across your channels.
Mark elaborated further on the pitfalls of disconnected systems:
Land-and-expand fails when billing, self-service, and CPQ systems are disconnected. You can’t easily offer “included-but-measured” features when three separate systems need to agree on the rules. Companies are switching to unified platforms because they literally couldn’t execute their pricing strategy with disconnected tools.
That’s all we’ve got for today. Thanks again to Mark and Tina for sharing their time and insights. For further reading, check out Nue’s Guide to Commit Burndown.
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 →