"But I Need More Than Six Tools" — The Most Common Objection Answered

This is the email I get most often after someone downloads the hex constraint PDF. It usually arrives within 24 hours of the download, and it usually starts the same way: "I read the hex guide and I like the concept, but my situation is different. I actually need more than six tools because..." What follows is a list. The list is always longer than six items. The items always sound reasonable. And in about 80% of cases, the list doesn't survive an honest audit.

I'm not dismissing the objection. It's the right question to ask. The hex constraint — limiting your active AI tool stack to six — sounds arbitrary if you haven't tested it. The point of this article isn't to convince you that six is a magic number. It's to walk through the process that reveals whether your 10 or 12 or 15 tools are all doing real work, or whether some of them are just... there.

Why It Feels True

The feeling that you need more than six tools is almost always genuine. You're not making it up. You look at your stack and you can point to each tool and name something it does. The subscription management tool tracks your spending. The note-taking app holds your research. The project management tool organizes your tasks. The second LLM subscription gives you a different perspective when you need it. The image generation tool makes graphics. The automation platform connects things. The dedicated writing assistant helps with copy. The transcription tool handles meetings. The code assistant helps with development. That's nine tools and none of them sound frivolous.

The problem is in the framing. "I can name what this tool does" is a different statement than "this tool produces output I can't get from any other tool in my stack." The first statement is almost always true. The second statement, applied rigorously, usually cuts the list in half.

Here's the pattern I see repeatedly: tools enter your stack one at a time, each solving a problem that was real at the moment of purchase. But stacks aren't designed — they accumulate. Nobody sat down and architected a 12-tool workflow from scratch. What happened is you needed image generation in March, so you got Midjourney. Then GPT-4o added image generation in June, and now you have two tools that generate images. You needed transcription for a project, so you got Otter.ai. Then Zoom added native transcription, and now you have two tools that transcribe. You tried Claude and liked it, but you kept ChatGPT "just in case," and now you have two general-purpose LLMs that you use for slightly different things but could, honestly, be consolidated into one.

Accumulation isn't the same as architecture. And the feeling of needing 12 tools is usually the feeling of having accumulated 12 tools and not wanting to admit that half of them are redundant.

The Audit That Settles It

There's one exercise that resolves this objection for most people, and it takes about 30 minutes. Get a piece of paper — or a spreadsheet, whatever you prefer — and list every AI tool you currently pay for or use regularly. For each one, write down:

1. The specific output it produces. Not what it does — what it produces. "Generates blog post first drafts" is an output. "AI-powered writing assistant" is a description. Be specific.
2. Whether any other tool in your list can produce that same output. Be honest. If Claude can generate blog post first drafts and Jasper can generate blog post first drafts, write that down.
3. The last time you actually used this tool to produce something you shipped or published or sent to someone. Not the last time you opened it. The last time it produced something real.

When I've done this exercise with readers — and I've now done it with a few dozen people through the Done-With-You sessions — the pattern is remarkably consistent. A list that starts at 10-15 tools usually audits down to 5-7 that are doing unique, current, active work. The rest fall into three categories: tools that overlap with something else in the stack, tools you haven't used in 30+ days, and tools you keep "just in case" for a use case that hasn't materialized.

The overlap category is the biggest. Most people with 12 AI tools have at least three pairs of tools doing substantially the same thing. Two LLMs. Two image generators. A note-taking app and a project management app that both have AI features doing the same summarization. An automation platform and a set of manual workflows in another tool that could be consolidated. The overlap isn't obvious because the tools have different interfaces and different branding, but the output — the actual thing they produce — is the same.

The "Just in Case" Trap

The tools that are hardest to cut are the ones you keep for insurance. "I don't use ChatGPT daily, but sometimes I need a second opinion on a Claude output." "I don't use Runway every week, but when I need video, I need video." "I don't use this transcription tool anymore, but I might need it for a meeting next month."

These tools survive by being cheap enough that the cancellation isn't worth the effort, and by triggering loss aversion every time you consider cutting them. Canceling a $20/month tool you use twice a month feels like giving something up. Keeping it feels like maintaining capability. The math says otherwise — $240/year for a tool you use 24 times is $10 per use, which is almost certainly more expensive than re-subscribing for the one month you actually need it — but the math doesn't feel as real as the fear of needing something you no longer have.

The fix is simple but requires a small act of courage: cancel the "just in case" tools and see if you actually miss them. Not whether you feel nervous about missing them — whether you actually encounter a situation where you need the tool and can't do the work without it. In my experience, and in the experience of most people I've worked with on this, the moment never comes. Or it comes once, and the cost of re-subscribing for one month is far less than the cost of maintaining the subscription for the 11 months you didn't use it.

The Genuine Exceptions

I said 80% of cases don't survive the audit. That means 20% do. Some workflows genuinely require more than six tools, and the hex constraint isn't a religion — it's a framework that works for most people most of the time.

The genuine exceptions share common characteristics. They involve multiple distinct output types — someone who produces text content, audio content, video content, and code is working across more mediums than someone who primarily writes. They involve non-overlapping capabilities — each tool does something categorically different, not a slightly different version of the same thing. And they involve frequent use — every tool gets used at least weekly, usually daily.

If your audit reveals seven or eight tools that each produce unique output, get used frequently, and don't overlap with each other — congratulations, your workflow is genuinely complex enough to exceed six. The hex was never about a specific number being sacred. It was about the discipline of auditing, the willingness to cut, and the recognition that constraint improves focus. If your honest number is eight, your hex is eight. The principle still holds: choose deliberately, not by accumulation.

But run the audit first. Most people who tell me they need 12 tools discover they need six. The objection dissolves not because I argued them out of it, but because they looked at their own stack honestly for the first time.

What to Do Now

Download your credit card statement or check your subscription management tool. List every AI tool. Run the three-question audit. Be honest with yourself — not with me, I'll never see your list. With yourself. The goal isn't to hit six. The goal is to know, for certain, which tools are working and which ones are coasting. Once you know that, the cutting is easy. It's the not-knowing that makes the stack grow.

This article is part of the Hex FAQ series at CustomClanker.

Related reading: The Hex Explained, The Hex in the Wild, One Year of AI Tools — What Survived, What Didn't