The Tools That Weren't Worth Building (Until Now)

Executive Summary: Custom software tools that solve specialized, repetitive problems used to require months of development time—making them uneconomical for all but the highest-value applications. AI-assisted development has cut that time by 80-90%, flipping the ROI calculation. Tools that weren't worth building are now obvious investments. This article examines one example—a wire bundle calculator—and explores what this shift means for technical manufacturing organizations.

‍

Here's a scenario that plays out dozens of times a week in aerospace manufacturing:

A production technician is reviewing an assembly drawing. They need to know: how thick will this wire bundle be, and what size braid do we need over it?

It's a straightforward question with a deceptively complex answer. You need the outer diameter of each wire (which varies by part number, manufacturer, and jacket material). You need to account for how circular cross-sections pack together geometrically. You need to reference NASA specs and industry standards for acceptable tolerances. You need to select the right braid size from what's actually available.

The technician knows one thing for certain: this is an engineering question.

So they send an email. "Can you calculate the bundle diameter for drawing XYZ-123? Need to order braid."

If the engineer gets to it in a few hours, that's great. More likely it's a day or two, because engineers are busy with higher-priority work than answering "quick" calculations.

The technician waits. The assembly waits. The schedule slips by a day.

This happens constantly. And until recently, there was no good solution.

Why the Solution Didn't Exist

You might think: "Surely there's a wire bundle calculator somewhere?"

There are a few. But they're either:

• Embedded in expensive CAD software most people don't have access to
• Built for general electrical work, not aerospace specifications
• Maintained as internal spreadsheets at large companies, never made public
• Generic enough to be useless for actual production decisions

Why didn't anyone build a good, publicly available, aerospace-spec wire bundle calculator?

Because the economics didn't make sense.

Building a tool like this properly means:

• Parsing datasheets from major wire manufacturers to get exact outer diameters
• Implementing the geometric packing calculations correctly
• Integrating NASA specs and industry standards
• Building a clean interface anyone can use
• Testing it against real-world scenarios
• Maintaining it as specs change

That's months of dedicated developer time. Maybe $50,000-$100,000 in fully-loaded costs.

For a tool that solves a 15-minute problem that happens 50 times a week? The math didn't work. So the problem persisted: engineers got interrupted constantly for "quick" calculations, production waited, and everyone knew it was inefficient but accepted it as the cost of doing business.

What Changed: AI-Assisted Development

The same AI capabilities we use to answer technical questions and automate workflows? They're also dramatically accelerating software development itself.

Tasks that used to take a developer days now take hours. Complex implementations that required weeks of careful coding can be prototyped in an afternoon and refined over a few days.

This changes what's economically viable to build.

Suddenly, that wire bundle calculator—the one that wasn't worth building for months of effort—becomes worth building for a week of effort.

The ROI calculation flips. The long tail of "specialized tools that would help but aren't worth it" becomes "specialized tools we should absolutely build."

The Wire Bundle Calculator: A Case Study

So we built it.

https://mold.modules.dev.xskel.com/calculator

Here's what it does:

You enter part numbers for the wires in your bundle. Actual manufacturer part numbers—M22759/16-20-9, M22759/11-22-0, whatever you're working with.

It looks up the exact specifications from our parsed database of manufacturer datasheets and relevant NASA/industry specs. Outer diameter, jacket material, insulation thickness.

It calculates bundle diameter using proper geometric packing algorithms, not approximations.

It recommends braid sizes that will actually fit over that bundle.

It takes 30 seconds. Not two days.

The Workflow Transformation

Before:

1. Production technician reviews drawing
2. Realizes they need bundle diameter calculation
3. Emails engineering
4. Waits 1-2 days
5. Gets answer
6. Orders braid
7. Proceeds with assembly

After:

1. Production technician reviews drawing
2. Opens calculator, enters part numbers
3. Gets answer in 30 seconds
4. Emails engineering: "Used the calculator, got these numbers. Ordering braid and proceeding unless you see an issue."
5. Proceeds with assembly

Notice what changed: the technician isn't blocked. Engineering reviews instead of calculates. The email shifts from "please do this for me" to "here's what I'm doing, flag if wrong."

The bottleneck disappears.

Beyond Human Access: Tools for AI Agents

Here's where it gets more powerful.

We didn't just build a website. We built it so AI agents can use it too.

When someone talks to our AI assistant and asks: "I need to know the bundle diameter for this wire list from drawing ABC-123," the AI doesn't try to calculate it. LLMs are terrible at math and you don't want them guessing at specifications.

Instead, the AI uses the calculator as a tool:

1. Extracts the wire part numbers from the drawing
2. Calls the bundle calculator with those part numbers
3. Gets back the definitive answer with sources
4. Explains the result to the user in plain language

The same tool serves both direct human use and AI-assisted workflows. That doubles its value without doubling its cost.

The Pattern Scales

Wire bundle calculations are just one example. Once you realize custom tooling is economically viable, you start seeing the possibilities everywhere:

Torque specifications - "What torque for this fastener in this material?" Requires looking up specs, cross-referencing material compatibility charts, checking installation notes. A tool could give you the answer with sources in seconds.

Connector compatibility - "What mating connector and crimp tool for part XYZ at 18 AWG?" Requires datasheet hunting and cross-checking specifications. A focused tool could provide instant answers with datasheet citations.

Heat shrink sizing - "What size heat shrink for this bundle at this temperature rating?" Complex calculation involving bundle diameter, shrink ratios, and temperature derating. A purpose-built calculator could handle it definitively.

Material cure schedules - "At what temperature and duration for this adhesive on this substrate?" Spec-heavy decisions that vary by manufacturer and application. A tool could pull from datasheets and provide the correct parameters.

Fill ratio calculations - "Will this connector accommodate these wire gauges?" Involves cross-sectional area calculations and derating factors. Easy to get wrong manually, straightforward for a specialized tool.

Each of these used to be "not worth building a tool for" because the development cost was too high relative to the problem. Each one now represents a viable investment because AI-assisted development has changed the economics.

The exoskeleton emerges from accumulating these capabilities. Individually, each tool saves 15 minutes here, an hour there. Collectively, they transform what your organization is capable of—not by replacing expertise, but by democratizing access to correct answers and freeing experts for higher-value work.

The Economics: Why This Matters

Let's be conservative with the wire bundle calculator example:

• Used to take 15 minutes of engineer time plus 1-2 days of waiting
• Happens approximately 50 times per week across the organization
• Now takes 30 seconds with no engineering time

Time saved: 12.5 hours of engineering time per week, plus eliminating days of wait time for production.

Annual impact: ~600 hours of engineering time freed up for complex work instead of routine calculations. Plus production moving faster without bottlenecks.

If the tool took a week to build and an hour a month to maintain? That's a 50:1 return on investment in year one alone.

This is the shift. Tools that previously couldn't justify their development cost now have overwhelming ROI.

Why Public, Not Hidden?

You might wonder: why make these tools publicly available instead of keeping them internal?

Because friction matters.

People need answers where they work. If the calculator requires logging into a special system, launching specific software, or remembering where it lives in your intranet—it won't get used.

A simple URL anyone can access from any device? That gets used. That becomes part of how work gets done.

Plus, making tools publicly available:

• Reduces your support burden (it's obvious how to use)
• Improves the tool through broader feedback
• Positions you as a capability builder in your industry
• Makes it easy to integrate with AI agents

The value is in having the tool and knowing how to use it in your workflows, not in keeping it secret.

The Bigger Vision: An Exoskeleton of Small Tools

Marc Andreessen famously said "software is eating the world" back in 2011. The thesis was that software economics would eventually reshape every industry.

He was right, but the timing was off. Writing quality software was expensive enough that only the highest-value applications justified the investment.

AI-assisted development changes that equation. Now software can eat the long tail—all those specialized, niche tools that would help specific industries or workflows but were never worth building.

At xSkel, this is our thesis: we can build an exoskeleton for organizations by creating dozens of small, focused tools that previously weren't economically viable.

Each tool is narrow. Each solves one specific problem. But together, they raise the floor—they establish a baseline capability that everyone in the organization has access to, regardless of their expertise level.

The wire bundle calculator is one brick. We're identifying and building more:

• Tools for technical specifications and standards compliance
• Tools for process calculations that require spec lookups
• Tools for procurement decisions based on component compatibility
• Tools for quality checks that cross-reference multiple data sources

Each one turns "ask an expert and wait" into "get the answer and verify."

That's the exoskeleton: lightweight supports that let everyone lift more with less strain.

What This Means for Your Organization

If you're running technical manufacturing operations, ask yourself:

What are the "quick" questions that eat up expert time in your organization?

Those 15-minute calculations that happen constantly. The specification lookups that require digging through PDFs. The compatibility checks that require institutional knowledge. The calculations that "someone should really automate" but never get prioritized.

Those are your candidates.

Each one used to require months of development effort to solve properly. Each one now might take a week or two.

The calculus has changed. The tools that weren't worth building? They're worth building now.

And once you build them—once you connect them to both direct human use and AI-assisted workflows—you create compound leverage. The tool serves multiple use cases, accelerates multiple workflows, and raises capabilities across multiple roles.

That's how you build a quality floor. Not with one massive system, but with an accumulation of small, focused capabilities that together transform what your organization can do.

Getting Started

You don't need to build everything at once. Start with one tool:

1. Identify the bottleneck - What "quick" question happens 20+ times per week?
2. Define the inputs and outputs - What information goes in, what answer comes out?
3. Find the authoritative sources - What specs, standards, or data define the correct answer?
4. Build the minimum viable tool - Clean inputs, definitive outputs, clear assumptions
5. Make it accessible - Simple interface for humans, clean API for AI agents
6. Measure the impact - How much time saved? How many bottlenecks eliminated?

Then build the next one. And the next.

Over time, these tools accumulate into your exoskeleton—the infrastructure that raises everyone's capabilities and establishes a floor of quality beneath which your organization cannot fall.

The economics changed. The tools that weren't worth building are worth building now.

‍