There's no shortage of new technology promising to change how SaaS companies operate. AI is moving quickly, product expectations are shifting, and the pressure to keep up is real.

But moving faster doesn't always mean making better decisions.

This month, I'm looking at three areas where SaaS leaders need to think carefully: when AI is worth the investment, how to build products that work with AI agents, and how to price embedded analytics as customer usage grows.

Each comes with opportunities, trade-offs, and decisions worth getting right.

Let's dive in.

QRVEY PRESENTS
A Product-Led Alliance Lightning Lecture
Qrvey
 
23
MISTAKES
Costly Mistakes SaaS Companies
Make With Embedded Analytics
WHEN
December 9, 2026
FORMAT
Online Lightning Lecture
Embedded analytics can make your SaaS product more valuable, but the wrong decisions can create expensive problems down the road.
Join Qrvey's CTO for a practical look at common embedded analytics mistakes and what SaaS teams should consider before making their next move.
 
YOUR SPEAKER
David Abramson
Chief Technology Officer, Qrvey
Learn what to avoid before it costs you.
RESERVE YOUR SPOT →

This Month’s Focus
Moving faster doesn't always mean making better decisions.
From deciding when AI is worth the investment to designing for AI agents and pricing embedded analytics, SaaS leaders face choices that can shape their products for years. The challenge is knowing where to invest, what to prioritize, and how to turn new capabilities into real customer value.

📰 Upcoming in this issue

  • The Hidden Cost of Chasing AI

  • Designing Features That AI Agents Can Actually Use

  • How to Package Embedded Analytics for Usage-Based Pricing

📈 Trending news

The Hidden Cost of Chasing AI

What if trying to catch up with AI is actually putting your SaaS company further behind?

I think that's a real risk right now.

There's enormous pressure on SaaS CEOs to demonstrate progress with AI. Competitors are announcing new capabilities, customers are asking about roadmaps, and nobody wants to be the company that missed the next major technology shift.

But the pressure to move quickly can lead to some very expensive decisions.

We experienced this firsthand at Qrvey.

Six months ago, if we had made some of the changes to our engineering processes that we've made today, I believe they would have failed. The AI tools simply weren't ready.

Today, we're using AI to improve how our teams develop software. We're shipping faster, improving quality, and delivering releases more consistently.

What changed wasn't our ambition. The technology became capable of solving problems we already knew we had.

And I think that's where many companies get the sequence wrong.

They start with AI and go looking for places to use it, rather than starting with a business problem and asking whether AI can help.

I've seen this happen with other technology shifts. Companies invest heavily in what they believe will be the next big thing, only to discover that customers aren't ready or the technology doesn't deliver what they expected.

If I were a SaaS CEO worried about falling behind, I'd ask three questions before committing significant resources:

  1. What problem are we actually trying to solve?

  2. Can AI meaningfully solve it better than our current approach?

  3. Is the technology ready to deliver that value today?

Moving too slowly has consequences. But so does investing too early in something that isn't ready.

Every AI investment comes at the expense of something else your company could be building.

And that's what makes getting the timing wrong so expensive. You don't just lose the money you invested. You lose the opportunity to build something that actually mattered.

Key Takeaways

1. Start with the business problem, not the AI technology. Identify the problem you're solving and whether AI offers a meaningful improvement over your current approach.

2. Prioritize technological readiness over competitive pressure. Evaluate whether AI is mature enough to deliver reliable results before committing engineering resources.

3. Consider the opportunity cost of every AI investment. Chasing AI initiatives that aren't ready diverts resources from product improvements that could deliver greater value today.

The Cost of Moving Too Soon
The pressure to adopt AI can make waiting feel risky. But investing before the technology is ready can be just as costly.
Why it matters: Every AI initiative uses engineering time, budget, and attention that could go toward something else. Before committing resources, SaaS leaders need to know what problem they're solving, whether AI offers a meaningful improvement, and if the technology can reliably deliver that value today.
The takeaway: Don't let competitive pressure set your AI roadmap. Let customer needs and technological readiness guide the investment.

Designing Features That AI Agents Can Actually Use

Software has traditionally been designed for two audiences.

People interact through screens, menus, and buttons. Developers work through APIs, code, and functions.

AI introduces a third kind of user: Agents.

An AI agent can interpret a person’s intent, invoke software capabilities, and coordinate multiple actions on their behalf.

This creates a new consideration for SaaS product teams: How should features be designed so AI agents can use them reliably?

Designing Features for Agents Changes the Rules
Traditional interfaces impose practical limits on feature design.

Every capability must be translated into something a person can see, understand, and operate. As features become more sophisticated, interfaces become harder to navigate.

Product teams often respond by simplifying capabilities, restricting them to power users, or leaving them out entirely.

An AI agent operates under different constraints.

A function might have dozens of parameters that would overwhelm a human interface. An agent can interpret the desired outcome, identify the relevant options, and invoke the function without exposing that complexity to the user.

But agents cannot reliably operate capabilities that were never designed for them to access.

Product teams need to consider how features are structured beneath the interface. This includes clear APIs, meaningful metadata, modular functions, explicit permissions, and predictable outputs.

Consider an agent designed for comparative analytics.

A user might ask it to compare revenue performance across regions. The agent needs to understand available measures, comparison methods, and relevant business definitions.

If the request is incomplete, it can ask whether to compare regions against the median or the highest performer.

The user provides the intent. The agent handles the complexity of executing it.

Keep People in Control
Designing features for AI agents doesn’t mean eliminating traditional interfaces.

Dashboards and direct controls remain valuable. People still need to inspect results, correct assumptions, and approve consequential actions.

The opportunity is to make software capabilities accessible to both people and agents, without requiring either to operate the same way.

Key Takeaways for SaaS Leaders

  1. Treat AI agents as another user of your software. Design features that agents can understand, invoke, and combine on behalf of people.

  2. Rethink the limitations of human interfaces. Agents can operate sophisticated capabilities without exposing every parameter or configuration to users.

  3. Make capabilities accessible and reliable. Clear APIs, structured metadata, permissions, and predictable outputs allow agents to work effectively.

  4. Preserve human oversight. Users must be able to inspect results, correct assumptions, and control consequential actions.

Pricing & Packaging Analytics: Part 3 of 3

How to Package Embedded Analytics for Usage-Based Pricing

This is the third 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 usage-based pricing model.  

For SaaS companies with usage-based pricing, monetizing embedded analytics requires a different approach than tiered pricing.

Tiered pricing monetizes access to capabilities. Usage-based pricing monetizes growth.

As customers process more transactions, manage more data, or handle more operational complexity, their analytical needs grow.

Embedded analytics should follow the same pattern.

Rather than charging separately for reports and dashboards, SaaS companies can monetize the advanced analytics capabilities customers need as their operations scale.

Let Customer Growth Drive Analytics Value

A company processing 1,000 transactions per month has different analytical needs than one processing a million.

Basic reporting may be sufficient at lower volumes. But as operations expand, more transactions, workflows, and exceptions create new challenges.

That's where advanced analytics becomes increasingly valuable.

Customers aren't paying for more reports. They're paying for capabilities that help them manage greater complexity.

Fit Analytics into Existing Usage-Based Pricing

Customers already understand that their costs increase as they use more of a product.

Analytics should reinforce that relationship, not introduce a separate pricing model.

Advanced capabilities can be incorporated into existing volume bands, usage commitments, or feature entitlements.

Consider a logistics SaaS platform that prices its subscriptions based on shipment volume.

All customers receive basic reporting on shipment volumes, delivery performance, and exceptions. As customers expand across more carriers, warehouses, and regions, managing delivery performance becomes more complex.

They may need self-service analytics to investigate bottlenecks, anomaly detection to identify unusual patterns, and predictive alerts to flag shipments at risk.

These capabilities become more valuable as operational complexity increases, creating opportunities to monetize advanced analytics within existing usage-based pricing tiers.

Keep Packaging Aligned with Customer Needs

Avoid introducing a separate analytics meter that complicates pricing.

Make advanced capabilities available when customers have a meaningful need for them. And revisit packaging regularly as once-premium capabilities become standard expectations.

The goal is to monetize the operational value analytics delivers as customers grow.

Key Takeaways

  1. Let customer growth drive analytics monetization. Package advanced capabilities around the operational challenges that emerge as usage increases.

  2. Build on your existing pricing model. Incorporate analytics into usage bands or entitlements rather than introducing a separate pricing structure.

  3. Match capabilities to customer maturity. Introduce advanced analytics when growing complexity makes its value clear.

  4. Revisit packaging as expectations evolve. Today's premium capabilities may become tomorrow's standard features.

Why It Matters

SaaS companies have more ways to build, improve, and monetize their products than ever before. That also means more opportunities to invest in the wrong things.

The common thread across this month's articles is making deliberate product decisions. Whether you're adopting AI, designing for agents, or packaging analytics, start with the customer problem and the value you're trying to deliver.

The technology will keep changing. Knowing what deserves your team's time and resources is what will help you stay ahead.

With Thanksgiving coming up next week, I also want to wish everyone celebrating a wonderful holiday. I hope you get some time to unplug, recharge, and enjoy it with family and friends.

Happy Thanksgiving!

Until next time,

Arman Eshraghi
The Embedded Intelligence Brief
Connect on LinkedIn

P.S. Are you making these 23 embedded analytics mistakes? Join Qrvey’s CTO on December 9 for a Lightning Lecture with Product-Led Alliance covering common mistakes SaaS companies make with embedded analytics and how to avoid them. Save your spot here.

Login or Subscribe to participate