The real measure of modern software productivity is not how much AI a team uses. It is how quickly and efficiently that team turns a genuine customer need into something valuable.

AI has dramatically shortened the distance between an idea and a working piece of software.

A developer can describe a feature and receive a functional implementation. An agent can work through a backlog of tasks. Teams can generate tests, documentation, prototypes, queries, designs, and code at a pace that would have seemed unrealistic only a few years ago.

That creates an obvious temptation: measure the AI activity itself.

Private real estate has long been a cornerstone of institutional portfolios, but individual investors have often had limited access to the asset class - or have had to invest through pooled funds and intermediaries.

Lightstone DIRECT offers a different approach. Accredited investors can invest directly in select real estate opportunities alongside Lightstone, a vertically integrated owner/operator with more than four decades of experience and a $12 billion real estate portfolio.

We invest 20% or more of the equity in every offering, aligning our interests with yours from day one. Our in-house team oversees acquisitions, asset management, and investment execution, with property operations managed by Lightstone and, where appropriate, experienced third-party operators.

No blind funds. No intermediaries. Just direct partnership.

Explore Opportunities

^{This communication is for informational purposes only and does not constitute an offer to sell or a solicitation of an offer to buy any securities. Any such offer or solicitation will be made exclusively through the definitive offering documents. All investments involve risk of loss, including the potential loss of principal. Past performance is not a guarantee of future results. Any targeted returns or projections are forward-looking statements, are based on current assumptions, and are not guarantees of future performance. Actual results may differ materially.}

How many pull requests were generated with AI? How many engineers use coding assistants? How many tokens were consumed? How much code was produced? How many AI-generated changes were merged?

Those numbers may be interesting.

But they can easily become a distraction from the thing that actually matters.

Did someone receive something useful, and how efficiently did the organization deliver it?

That question becomes even more important as AI makes producing software cheaper and faster. When execution accelerates, choosing the right work becomes a bigger part of the problem.

Software organizations have always struggled with measuring productivity.

Lines of code were once treated as an indication of output. Then came story points, tickets completed, deployment frequency, pull requests, and other measurements intended to reveal whether teams were becoming more effective.

AI introduces an entirely new collection of possible measurements.

A team can proudly announce that most of its pull requests now involve AI. Another organization can point to dramatic increases in generated code. A development platform might show which prompts produce the highest merge rate.

But none of these numbers directly answers the question a customer cares about.

Suppose an AI assistant helps a team produce twice as many pull requests. If those changes do not improve the product, solve customer problems, reduce operational costs, or create meaningful business outcomes, the organization has increased activity without necessarily increasing value.

That distinction matters because AI makes output abundant.

When writing software required significant human effort, the cost of producing more work naturally limited how much could be created. AI weakens that constraint. Teams can now generate more possibilities than they have time to evaluate.

The bottleneck therefore starts moving upstream.

It becomes less about "Can the team build this?" and more about "Should the team build this at all?"

Tip: When evaluating an AI-enabled workflow, start with the customer or business outcome and work backward. Only then decide which activity metrics are useful for explaining how that outcome was achieved.

There is nothing wrong with collecting more engineering data.

In fact, better visibility can help teams understand where work slows down, where bottlenecks appear, and which practices are actually improving delivery.

The danger comes when more detailed data creates the illusion of greater certainty.

A dashboard containing dozens of AI metrics can look sophisticated while still failing to answer the most important question.

Imagine seeing that an AI coding assistant generated thousands of suggestions, that engineers accepted a large percentage of them, and that pull-request throughput increased substantially.

That sounds impressive.

Now add one missing number: customers are still waiting the same amount of time for important problems to be solved.

The impressive dashboard suddenly looks much less useful.

The right metric depends on where you sit in the organization and what decisions you can actually influence.

A company leader may care about the financial return from a major technology initiative. A product organization may care about customer adoption or time to deliver a meaningful capability. An engineering team may need more granular measurements to understand where work is getting stuck.

There is no single metric that should replace everything else.

The better approach is to connect measurements across levels so they all point toward the same underlying outcome.

Tip: Keep detailed engineering metrics, but make sure every important metric has a clear connection to a decision or outcome. If nobody knows what action a number should trigger, question whether it belongs on the dashboard.

For software teams, one particularly useful concept is cycle time.

Instead of asking how much work was produced, ask how long it took to turn a meaningful customer need into a delivered result.

That changes the conversation.

Suppose customers repeatedly struggle to find the right product quickly. The valuable outcome is not the number of tickets completed or AI-generated pull requests.

The valuable outcome is giving those customers a better way to find what they need.

Cycle time asks how long it takes to move from recognizing that need to actually delivering the improvement.

That makes the metric much harder to detach from reality.

Customers do not care whether an AI agent generated 300 lines of code in 30 seconds. They care whether the feature works and whether the problem was solved.

Cycle time can also be examined at different levels without changing the fundamental idea.

At the organizational level, it could measure the time between identifying a customer need and delivering the resulting capability.

At the team level, it could measure the time between a requirement entering development and reaching customers.

At the engineering level, it could examine how long individual changes take to move through the system.

These measurements are different in scope, but they describe the same basic question:

How quickly does useful change move through the organization?

Tip: Use cycle time at the level your team can influence. Avoid forcing an engineering team to optimize a business outcome it cannot directly control, but make sure its local measurements still connect to the larger delivery chain.

There is an important limitation to focusing heavily on speed.

A team can deliver something quickly and still make a terrible decision.

AI makes this easier because the cost of experimentation has fallen. Building a prototype that once required several days might now take several hours. Producing multiple approaches can be nearly effortless.

That is useful.

But faster construction does not guarantee that the resulting product deserves to exist.

A company could spend enormous resources rapidly building features that customers barely use. The development cycle may look excellent while the economic result remains poor.

This is why time alone cannot define value.

The better question is the relationship between the investment required and the outcome produced.

And the word "investment" is becoming broader.

It includes engineering and product time, infrastructure, tooling, AI services, model usage, evaluation, monitoring, and the organizational effort required to maintain what gets built.

AI does not eliminate the cost of software.

It changes the composition of that cost.

Tip: Whenever a new AI workflow makes development dramatically faster, ask what additional costs it introduces. Faster execution is valuable only when the resulting outcome justifies the total investment.

For years, software organizations could hide inefficient development behind a simple assumption: build something useful and, eventually, its economics will work.

That assumption was always risky.

A project could take years and consume millions while still producing disappointing returns. Yet long development cycles and substantial investment budgets sometimes made those problems difficult to see immediately.

AI makes the mismatch harder to ignore.

The cost of producing software is changing because part of the work can now be performed by agents and models. But that does not mean every additional piece of software becomes economically attractive.

In fact, cheaper production can encourage organizations to build too much.

If creating another feature becomes inexpensive, the natural response may be to create more features. But every feature still creates maintenance obligations, support requirements, operational complexity, security considerations, and user-experience consequences.

The organization can therefore become extremely efficient at producing things that it should never have built.

That is why selection becomes as important as execution.

When the cost of building falls, the cost of choosing badly becomes relatively more significant.

AI adoption often begins with a technology question:

"Which AI tools should the team use?"

That is a reasonable starting point, but it is not where the conversation should end.

A stronger question is:

"What would allow this organization to deliver meaningful outcomes faster and with less unnecessary investment?"

Now AI becomes one possible answer among many. Maybe an AI coding assistant reduces implementation time. Maybe automated testing removes a bottleneck. Maybe better requirements prevent unnecessary rework. Maybe smaller releases allow the team to validate ideas earlier. Maybe an architectural simplification removes months of future friction. Maybe the biggest improvement has nothing to do with AI at all.

This is important because AI can become another optimization target that distracts from the system as a whole.

The goal is not to maximize AI usage.

The goal is to improve the economics and speed of delivering useful outcomes.

Tip: Treat AI as an input to your delivery system rather than the objective of the system. Measure whether it improves outcomes, cost, quality, or speed—not simply whether people are using it.

HubSpot Academy's Digital Marketing Certification covers SEO, email, paid ads, social, and AI — in just over 3 hours, at zero cost. Join 200,000+ professionals who have advanced their career with HubSpot Academy. Get started today.

Get Certified Free

There is an interesting side effect of measuring cycle time.

Teams naturally begin looking for ways to make work smaller.

Instead of spending six months building one enormous capability, they may look for a way to deliver the first useful slice in a few weeks.

That can improve more than speed.

Smaller changes are generally easier to test, review, deploy, and reverse. They also provide earlier feedback about whether the underlying assumption was correct.

AI makes this approach increasingly practical because teams can prototype and iterate much more quickly.

But the important point is that the measurement does not have to encourage artificial activity.

If the goal remains meaningful customer value, smaller batches become useful because they allow value to reach users sooner and reduce the amount of work committed before feedback arrives.

This creates a healthier relationship between speed and value.

Tip: When cycle time looks too long, don't immediately ask the team to work faster. First ask whether the work can be divided into smaller valuable increments.

One of the biggest changes AI introduces is the ability to spend resources in ways that are less visible than traditional engineering labor.

Token usage can accumulate quickly. Agent loops can run repeatedly. Automated systems can perform work without requiring someone to sit in front of a keyboard.

That makes measurement of investment more important.

A team might save hours of engineering time while dramatically increasing model usage. Another might increase AI spending but reduce the time required to deliver a major capability.

Neither outcome can be judged from the AI bill alone.

The meaningful question is whether the combined investment produced enough value to justify itself.

This is why time and money should be considered together.

If a process costs more but delivers substantially greater value much sooner, it may be worthwhile.

If a process saves engineering time but produces little that customers actually need, the efficiency is mostly cosmetic.

Tip: Track AI costs alongside delivery time and outcomes. A lower engineering effort is not necessarily an improvement if the total cost rises without producing a better result.

There is a reason the idea of productivity can become so confusing in software.

The industry often measures the things that are easiest to count.

Pull requests are visible. Tokens are visible. Deployments are visible. Tickets are visible. Customer value is harder.

It requires understanding whether the problem being solved actually matters, whether the solution works, whether customers use it, and whether the organization delivered it at a reasonable cost.

That makes value less convenient to measure.

It does not make it less important.

AI is simply making this distinction impossible to ignore because the technology can dramatically increase the amount of work an organization is capable of producing.

When production capacity expands, the organization needs stronger judgment about what deserves that capacity.

The winning workflow is therefore unlikely to be the one with the highest AI adoption percentage.

It will be the one that consistently turns limited resources—people, time, money, and increasingly AI capacity—into useful outcomes.

For you, the practical takeaway is surprisingly simple.

You do not need another dashboard celebrating how many AI-generated changes were produced this month.

You need a clearer connection between need, investment, time, delivery, and outcome.

AI can shorten the path between an idea and implementation. That is a major advantage.

But the shorter that path becomes, the more important it is to ensure that the destination is worth reaching.

The organizations that figure this out will not necessarily be the ones using the most AI.

They will be the ones that use AI to make meaningful work faster, cheaper, more focused, and easier to validate—while refusing to confuse technological activity with actual progress.

Tip: The next time an AI metric looks impressive, trace it all the way to the customer. If the connection disappears somewhere along the way, you may be measuring the machinery instead of the value.

What’s your next spark? A new platform engineering skill? A bold pitch? A team ready to rise? Share your ideas or challenges at Tiny Big Spark. Let’s build your pyramid—together.

That’s it!

Keep innovating and stay inspired!

If you think your colleagues and friends would find this content valuable, we’d love it if you shared our newsletter with them!

As the world becomes increasingly digital, this question will be on the minds of millions of people seeking new income streams in 2026.

The answer is—Absolutely!

Thank you for taking the time to read today’s email! Your support allows me to send out this newsletter for free every day. 

 What do you think for today’s episode? Please provide your feedback in the poll below.

Share the newsletter with your friends and colleagues if you find it valuable.

Disclaimer: The "Tiny Big Spark" newsletter is for informational and educational purposes only, not a substitute for professional advice, including financial, legal, medical, or technical. We strive for accuracy but make no guarantees about the completeness or reliability of the information provided. Any reliance on this information is at your own risk. The views expressed are those of the authors and do not reflect any organization's official position. This newsletter may link to external sites we don't control; we do not endorse their content. We are not liable for any losses or damages from using this information.