What Do You Do With the Hours You Get Back?

Almost every conversation our team has about AI in development starts with cost. How much will this save us? What's the discount?

It's a fair question. It's also the least valuable thing you can do with the capacity.

There are two better answers, and one honest counterargument that deserves a real response. Let's take them in order.

Start With Ford, Because the Mistake Is the Same

When Henry Ford put the moving assembly line into Highland Park in 1913, the obvious play was labor savings. Build the same number of cars with fewer people. Bank the difference.

That is not what he did. He held the workforce, cut the price of a Model T from around $850 to under $300 over the following decade, and sold into a market that had not existed before because nobody at that price point had been able to buy a car. He also doubled wages to five dollars a day, which looked insane at the time and turned out to be the cheapest labor-retention decision in industrial history.

The efficiency gain was real. The value came from what he chose to do with it.

This is how our team thinks when brands ask us about AI in delivery. The efficiency gain from agentic development is measurable: 238% over human-only delivery in our work, with an 82% agent success rate on assigned tasks. That number is not the interesting part. The interesting part is what you decide it buys.

1. Redeploy the Hours Into Work That Compounds

The first path is to keep your development spend flat and take the reclaimed hours back into your own team's capacity.

Here is what that looks like in practice. Your merchandising lead stops waiting three weeks for a promo module and starts testing merchandising strategy. Your analytics person stops filing tickets for instrumentation and starts reading the data the instrumentation produces. Your content team stops planning the editorial calendar around development availability and starts planning it around the business.

Every one of those shifts moves a person from managing a constraint to doing the work you hired them for. That is not a soft benefit. A merchandiser running four tests a quarter instead of one is worth substantially more than the ticket queue they were previously administering.

The compounding is the part people miss. Testing produces learning. Learning improves the next test. A team that gets four cycles a year instead of one is not four times better, it is further ahead than that, because each cycle starts from a better position than the last.

2. Or Buy More Development for the Same Budget

The second path is to hold your budget flat and take the gain as volume instead.

This is the right call for brands carrying real technical debt, or brands where development velocity is the thing standing between them and a roadmap they already believe in. You are not spending less. You are getting materially more shipped for what you were already paying, which turns delivery speed into a competitive position rather than a line item.

It also helps to treat this path as a progression rather than a one-time decision. Anyone who worked through the digital transformation era remembers the maturity curve, because nearly every strategy deck had one. You placed an organization on the curve, named the stage it needed to reach next, and scoped the work to move it one stage closer. Nobody expected a single project to carry them to the end of it.

AI maturity in delivery works the same way, and in our experience most enterprise ecommerce brands start from the same place. 73% of organizations cite development capacity as their primary limitation to growth, with backlogs averaging 14 months. That is not a budget problem. Those teams have already decided the work is worth funding. They cannot get it built.

If your backlog is the binding constraint on your roadmap, taking the gain as savings is choosing to stay constrained in order to spend slightly less staying constrained.

So the next stage is straightforward: use the added capacity to clear the backlog and pay down the technical debt. Nobody pictures debt reduction when they imagine what AI will do for them, but every stage after this one depends on it.

3. The Tradeoff This Actually Dissolves

Once the backlog is under control, the question changes from how to catch up to what to build next. This is the stage where the reinvestment question answers itself.

We hear some version of the same problem constantly. Budget pressure forces teams to choose between must-do work and the work that advances strategic business objectives. Platform upgrades, security patches, and compliance work are not optional and produce nothing a customer will ever notice. Most teams fund the must-do work, defer everything else, and repeat that decision until the strategic work stops being proposed.

That tradeoff exists because capacity is fixed and the backlog absorbs most of it. Clear the backlog and the must-do work shrinks back to routine maintenance. The Drupal core upgrade stops consuming the quarter, and the capacity that used to disappear into catch-up goes toward conversion work, new customer experiences, and the initiatives leadership set out to fund in the first place.

Our team would rather a client spend their reclaimed capacity here than anywhere else, because this is the decision that quietly determines whether a digital program compounds or just persists. As with any maturity curve, the only way to this stage is through the one before it.

4. The Counterargument, Honestly

There are several good arguments for just taking the savings, and we want to give them a fair hearing rather than wave at them.

If your organization is under genuine financial pressure, reduced spend is not a failure of imagination, it is responsible management. If your roadmap is genuinely complete and your backlog is genuinely short, more capacity has nowhere useful to go. If you are mid-acquisition or planning a replatform in six months, holding cash and deferring investment is defensible.

Those situations are real. They are also less common than the number of brands who default to savings because it is the easiest thing to explain to a CFO.

The honest test is whether your backlog is empty (is it ever really?). If it is, take the savings and we'll tell you the same thing. If it is fourteen months deep, the savings are a rounding error against the cost of everything in that backlog continuing not to ship.

5. What This Means for How You Structure the Engagement

This is where agentic development connects to how you think about ongoing work generally. We have written before that the launch is the starting line, not the finish line. The reinvestment question is the same question asked earlier: what do you do with capacity once you have it, and are you spending it on improvement or on maintenance?

It is wise to heed the words of the late, great Charlie Munger, Berkshire Hathaway's vice chairman and longtime business partner of Warren Buffett: "Start with the end in mind." What outcomes from the website move the strategic objectives of the business forward? Structure your agentic development engagement to maximize those outcomes.

The Kicker

Ford did not get rich by building the same car with fewer people. He got rich by deciding what the efficiency was for.

Go look at your backlog. If the answer to what you would do with more capacity is obvious within thirty seconds, you already know which path you are on. Take the volume.

If you have to think about it for a while, that is worth knowing too, and it is a better conversation to have before an engagement than eighteen months into one.