Will AI commoditize software? LBM dealers will buy on trust

AI will commoditize most software features, though not the systems that hold a dealer's inventory, accounts and payments. Retool's 2026 survey found 35% of teams had already replaced a SaaS tool with custom software. When every vendor demos the same AI, lumber and building material dealers will choose on price, service, setup effort and trust.

Lumber is a commodity, and no dealer I know picks a supplier because the 2x4 is more inspired. Software is about to be bought the same way.

It cut the time to field a competing SaaS product from 18-24 months to 1-3 months, by one analyst's account, and let non-developers build internal tools at all.

Two sources put numbers on it. Retool's 2026 Build vs. Buy report surveyed 817 of its own customers and builders in late 2025. It found 35% had replaced at least one SaaS tool with custom software, 78% expect to build more internal tools in 2026, and 60% had built software outside IT oversight in the past year. The most-replaced categories were workflow automations and internal admin tools, then CRMs, BI, project management and customer support. Retool says it surveyed its own customers and builders, so the sample leans toward people who already build.

Andrew Stroup's April 2026 essay argues the timeline to compete in unregulated SaaS fell from 18-24 months to 1-3 months. It is one person's estimate, not a measurement. His claim is that payments infrastructure barely moved, and that the moat now sits in money movement, compliance and infrastructure embed.

Read both with that discount. The direction is hard to argue with.

A commodity is bought on price, on availability and on whether you trust the supplier to show up.

Dealers know this game. Dimensional lumber looks the same across suppliers, so the value lives in the delivery, the credit terms, the person who picks up the phone when a truck is late. The supplier you keep is the one who was there when something went wrong.

My view is that software enters the same game. When capability stops separating vendors, customers pick on what matters to them: a fair price, the service and relationship they want, and the path of least resistance into their operation. It comes back to relationships and trust, as it does whenever something rare becomes common.

Jennifer Courchaine of Vehlo said it plainly in a July 2026 ChurnZero discussion: "Decent software is going to become table stakes. Where you can differentiate is the service you provide."

For thin internal tools, increasingly yes. For systems of record, no, and the reason is integration and compliance rather than code.

A quote-request form, an internal dashboard or a routing helper is the kind of workflow tool Retool's respondents are replacing SaaS with. If AI lets a dealer's operations manager build one in an afternoon, a vendor charging for that feature has a problem.

Inventory, accounting, tax and payments are different. They have to agree with your ERP to the penny, survive audits and keep running after the person who built them leaves. Stroup's list of what resists is money movement, compliance burden and infrastructure embed, and an ERP-connected storefront sits on that last one.

The honest rule: build what you can describe in a paragraph and replace in a day. Buy anything your accountant, auditor or card processor will ask about.

ToolSwift is built to pass the path-of-least-resistance test, which is also why you should check my claims against the integrations page.

ToolSwift connects to sixteen named ERP and accounting systems, including Epicor (BisTrack, Eagle, Prophet 21, Eclipse), ECI (Spruce, Deacom), DMSi (Agility, Frameworks), QuickBooks and NetSuite, and builds connectors for others. It has distributor catalog integrations with Orgill, Do it Best and Emery Jensen. It is API-first, and the retail storefront and the contractor dashboard share one catalog, so a dealer maintains products once. The full list is on the integrations page.

The Takeoff Agent is live. It turns a contractor's voice note, photo, scan, takeoff or file into a draft quote that staff review before it goes out. It is a narrow, workflow-specific agent, which matches Marshall's advice to match AI to specific workflows. For the wider picture of AI inside a B2B store, see how AI levels the playing field for vendors.

I should be honest about what is not on that list. Features are the part of my pitch AI makes easiest to copy. Service and trust are the parts I have to earn dealer by dealer.

Run every vendor through the same five questions, and write the answers down.

Question five is the useful one. A vendor who says "build that yourself" is easier to trust on the rest.

My prediction: by the end of 2027, no dealer I talk to will rank vendors on AI features, and the ones who win will be the ones whose customers will take their call.

AI will commoditize features, not systems of record. Retool's survey found 35% of respondents had replaced a SaaS tool with custom software, mostly workflow and admin tools. Software tied to inventory, accounting, payments and compliance is harder to replace, because integration and trust are the work.

Not yet, but the features inside it are heading that way. When rivals can copy a feature in 1-3 months, as Stroup argues for unregulated SaaS, buyers fall back on price, service and trust. That is how dealers already buy lumber.

Test the four things AI does not copy: total price, service after the sale, how it connects to your ERP by name, and the vendor's track record in LBM. Ask each vendor to run their AI on your own catalog and quotes.

Build thin internal tools that you can describe in a paragraph. Buy anything that must reconcile with your ERP, tax, payments or audits. The Retool data shows dealers are not alone in building; it also shows the sample is Retool's customers and builders.