AI is changing how SaaS companies build products, but building faster doesn't automatically mean building better.
In this edition, we're looking at what happens when AI-generated prototypes reach production, why traditional analytics tools have been harder to use than they should be, and how SaaS companies can price analytics without limiting adoption.
We'll also share five lessons from our recent panel discussion with SaaS leaders on what works, what breaks, and what may be better to buy than build.
|
📰 Upcoming in this issue
Top 5 Takeaways from the Panel Discussion, “Building SaaS with AI: What works, what breaks, and what not to build.”
AI and the Five-Click Ceiling
How to Package Embedded Analytics for Seat-Based Pricing
📈 Trending news
AI Writes More Code. So Why Aren’t Teams Shipping More?
Is Your AI Product Just Someone Else's Feature?
Your Biggest Engineering Cost Isn't Cloud Computing
Top 5 Takeaways from the Panel Discussion, “Building SaaS with AI: What works, what breaks, and what not to build.”
We had a panel discussion to hear from SaaS leaders about where AI coding tools are creating real leverage in product development, and what happens when teams try to turn those fast prototypes into products that are ready for customers.
We gained different perspectives on where AI helps, where it breaks down, and what’s actually worth building.
The panel’s central takeaway: AI can accelerate development, but SaaS companies still own everything required to make the result secure, reliable, maintainable, and valuable to customers. The discussion highlighted these practical lessons for SaaS leaders.
AI amplifies your existing development process.
Strong architecture and disciplined engineering give AI a foundation to accelerate. Weak processes can produce more code, more complexity, and more rework faster. Improving the process matters as much as adopting the tools.Every feature creates a long-term ownership commitment.
The first build is only part of the cost. Maintenance, support, security, documentation, and future enhancements continue throughout the product’s life. Unofficial tools become especially risky when their creator leaves and no one understands how to maintain them.Human accountability stays with the SaaS company.
Someone must understand the system, explain its behavior, and restore service when it fails. AI-generated code and third-party components both require clear ownership; customers still hold your company responsible for the complete experience.Evaluate embedded analytics as a complete layer and not a capability.
Generating a few charts can hide the work underneath: data integration, semantic context, governance, security, and ongoing maintenance. A build decision needs to account for those layers and whether owning them fits your company’s core focus.Use faster prototyping to learn from customers more often.
Interactive prototypes make it easier to gather specific feedback before committing engineering resources. Panelists described small customer advisory groups, industry-focused sessions, and engineers working directly with customers to identify needs and bring recurring patterns back into the product.
The biggest mistake a SaaS company can make in the AI age is applying fast prototyping techniques to product development.
I invite you to watch the panel discussion here on qrvey.com.
Special thanks to our panelists: Bill Bice, CEO of Zebraworks, and Mahesh Rajasekharan, President and CEO, of Cleo. And thank you to Bill Hostmann, VP and Research Fellow at Dresner Advisory for sharing build vs. buy insights across their AI Application Framework.
AI and the Five-Click Ceiling
Business intelligence promised to make analytics accessible to everyone. But as platforms became more powerful, they also became harder to use.
This created the Five-Click Ceiling.
It's a rule of thumb: tasks requiring fewer than three clicks tend to see broad adoption, while those requiring more than five often go unused.
For years, BI vendors had to choose between powerful analytics and simple interfaces. Advanced features existed, but accessing them often required complicated menus, filters, and formulas.
The Problem Was Never the User
Business users usually know what they want to understand.
A sales leader might ask why one region is underperforming. An executive might want to know how much revenue could increase if every business line performed as well as its strongest peer.
These questions are easy to express but difficult to execute in traditional BI software.
Users must translate their questions into dimensions, measures, filters, calculations, and visualizations.
The challenge isn't understanding the analysis. It's getting the software to produce it.
AI Removes the Barrier
AI changes how people interact with analytics.
Instead of navigating multiple menus, users can simply ask:
Show me the revenue opportunity if every region performed at the median.
AI can translate that question into the necessary analytical steps, while the user focuses on interpreting the results and making decisions.
Traditional dashboards will still have their place, especially for routine tasks. But sophisticated analysis no longer needs to be limited by what fits into a simple interface.
The real value of AI in business intelligence isn't replacing human thinking. It's removing the software complexity that gets in its way.
Key Takeaways
The Five-Click Ceiling has limited BI adoption. When analytics tasks require too many steps, business users are less likely to use advanced capabilities, even when those capabilities could deliver valuable insights.
The barrier to advanced analytics is software complexity, not user ability. Business users often know what questions they want answered but struggle to translate those questions into the filters, formulas, and configurations traditional BI tools require.
AI removes the trade-off between analytical power and ease of use. By translating natural-language questions into analytical operations, AI makes sophisticated analysis accessible without requiring users to master complex software.
The future of BI is focused on outcomes, not interactions. AI enables users to focus on asking better business questions and interpreting results, while the software handles the complexity of producing answers.
Pricing & Packaging Analytics: Part 2 of 3
How to Package Embedded Analytics for Seat-Based Pricing
This is the second how-to article, in a series of three, that covers different pricing and packaging strategies SaaS leaders can use to monetize their analytics offering. This month we’ll focus on the seat-based pricing model – and why it still works.
Seat-based pricing works when analytics permissions reflect what users actually do, rather than simply who can access a dashboard.
The challenge is that analytics becomes more valuable when more people use it. Charging every viewer can discourage adoption and limit how widely insights are shared.
A better approach is to separate basic access from advanced responsibilities.
Standard users should be able to view dashboards, monitor KPIs, and use insights in their daily work. Premium users can unlock capabilities such as:
Creating and publishing custom dashboards
Defining advanced metrics and calculations
Managing data sources and permissions
Configuring alerts and automation
Using advanced or AI-assisted analytics
For example, a customer success platform could include health scores and renewal insights in standard seats while reserving custom reporting and analytics administration for premium roles.
The key is to charge for advanced capabilities without making everyday insights harder to access.
This protects adoption while giving customers clear reasons to upgrade.
Key Takeaways
Seat-based pricing should encourage analytics adoption, not restrict it. Charging every user to view dashboards and reports can discourage customers from sharing insights, limiting the value of embedded analytics.
Monetize advanced capabilities rather than basic access. Standard seats can include access to insights, while premium seats unlock dashboard creation, custom metrics, automation, and governance capabilities.
Align analytics pricing with existing user roles. Packaging analytics around familiar roles, such as standard users, managers, and administrators, simplifies purchasing decisions and makes premium capabilities easier to justify.
Keep pricing simple and tied to meaningful value. Avoid creating too many analytics roles or charging for minor feature differences. Premium seats should offer distinct capabilities that justify the additional investment while supporting broad analytics adoption.
Why It Matters
AI is giving SaaS companies the ability to build and test ideas faster than ever. But speed alone doesn't determine whether a product succeeds.
The real challenge is deciding what deserves to be built, making advanced capabilities easier for customers to use, and finding pricing models that reflect the value they deliver.
As we discussed in our recent panel, every feature comes with a long-term responsibility. The companies that get this right will be the ones that focus on solving real customer problems, not simply adding more features.
Until next time,
Arman Eshraghi
The Embedded Intelligence Brief
Connect on LinkedIn
P.S. Missed our Building SaaS with AI panel? Hear SaaS leaders discuss what works, what breaks between prototype and production, and how to decide what’s worth building yourself. Watch the on-demand discussion here.