Education Centre Guides & Checklists
Build vs buy: when to buy software, connect your tools or build your own
Build vs buy for growing businesses. When to buy software, connect your tools with automation, or build your own, plus a decision checklist.
At a glance
Build vs buy is the choice between buying off-the-shelf software, connecting the tools you already have with automation, or building a custom app, automation or AI agent. AI coding tools have made custom builds faster and cheaper since 2024, but buying or connecting is still the right answer for most standard problems. Use the checklist below to decide, then scope it properly before you commit.
- Options
- Buy, connect or build
- SaaS licences used
- 54% on average (Zylo, 2026)
- Replaced SaaS with a build
- 35% of teams (Retool, 2026)
- Promptwise starts with
- Scoping the whole business
Build vs buy is the decision every growing business hits when a tool doesn't quite fit: do you buy software off the shelf, connect the tools you already have, or build something of your own? There are really three options, not two, and the middle one is the one most people skip.
The aim is to pick the right option for each problem, rather than one rule for everything.
The three options at a glance
| Buy off the shelf | Connect your existing tools | Custom build | |
|---|---|---|---|
| When it fits | The problem is common and the process is standard (CRM, email, accounting, e-signature) | The tools work, but data and tasks fall between them | The process is core to how you win, or nothing on the market fits without heavy workarounds |
| Examples | HubSpot, Pipedrive, Xero | Zapier, Make, n8n, Clay for data | Internal apps, client portals, AI agents, custom workflows built with tools like Claude Code |
| Cost shape | Monthly per seat or per contact, rising as you grow | Low monthly fee plus usage (tasks, operations or credits) | Upfront build cost, then hosting and ongoing maintenance |
| Time to value | Days | Days to a few weeks | Weeks, with a working prototype often much sooner |
| Main risks | Seat creep, workarounds, paying for features you never use | Fragile chains of automations nobody owns or documents | Maintenance, security and who owns the code once it's built |
A good rule of thumb: buy for the standard stuff, connect to close the gaps, and build only where it gives you an edge or saves real money.
A decision checklist
Ask these before you sign up for anything or brief a build:
- Is this process standard or a point of difference? If every business in your category does it the same way, buy. If the way you do it is part of why clients choose you, building may be worth it.
- Does a tool already do 80% of it? If so, buy it and accept the other 20%, or close the gap with an automation.
- Is the real problem the gap between tools? Copying data by hand from one system to another is a connection problem, not a software problem.
- What will it cost at your size in two years? Price the plan you'll need with your future headcount, contacts and usage, not today's. The stack calculator helps here.
- How many workarounds are you running today? Spreadsheets, manual exports and "just remember to" steps are a sign the tool doesn't fit.
- Who will own it after launch? Every automation and every build needs a named owner who fixes it when it breaks.
- What data and security rules apply? Client data, payments and personal information raise the bar for anything you build or connect.
- Can you test it small first? A small prototype or a single workflow tells you more than a six-month plan.
Mostly "standard"? Buy. Mostly "the gap"? Connect. Mostly "point of difference" and "nothing fits"? Scope a build.
The hidden costs of buying
Off-the-shelf software looks cheap on the pricing page. The costs show up later.
- Seat creep. Per-seat pricing scales with your team, and the features you actually want are often a tier or two up. Zylo's 2026 SaaS Management Index found the average enterprise uses only 54% of its SaaS licences.
- Workarounds. When a tool almost fits, people build spreadsheets and manual steps around it. That time never shows up on an invoice, but you pay for it every week.
- Tool sprawl. The same Zylo report found the average organisation has 305 apps and adds about nine new ones a month. Every tool is another subscription, another login and another place for data to leak, which is why I always say start small and let use cases build the stack.
None of this means don't buy. It means buy deliberately and cancel what nobody uses. My own picks are in Adam's stack.
The hidden costs of building
Custom builds have their own traps, and they're easier to underestimate.
- Maintenance. Software isn't finished at launch. APIs change, tools you connect to update, bugs appear and your process evolves. Budget for ongoing care, not just the build.
- Ownership. Who holds the code, the hosting accounts and the logins? Get this in writing. If the person who built it leaves, you need to be able to keep running it.
- Security. A tool you build is only as secure as the way it was built and hosted.
- Scope creep. "While we're at it" is how a two-week build becomes a three-month one. Start with the smallest version that solves the problem.
- Shadow IT. When building gets easy, people build without telling anyone. Retool's 2026 Build vs. Buy Report, a survey of 817 builders, found 60% had built software outside IT oversight in the past year.
How AI coding tools have changed the economics since 2024
The biggest shift is that building has become much faster and cheaper. AI coding tools like Claude Code can write, test and change working software from plain-English instructions.
- Adoption is now mainstream. The Stack Overflow Developer Survey 2025 found 84% of developers use or plan to use AI tools, up from 76% the year before, and 51% of professional developers use them daily.
- AI is doing the work, not just advising. Anthropic's Economic Index analysis of 500,000 coding interactions (2025) classed 79% of Claude Code conversations as automation, where the AI carries out the task directly.
- Businesses are acting on it. Retool's 2026 Build vs. Buy Report found 35% of teams have already replaced at least one SaaS tool with a custom build, and 78% expect to build more custom internal tools in 2026. Workflow automations and internal admin tools are the most common replacements.
But it isn't magic. In a 2025 randomised trial, METR found experienced open-source developers took 19% longer on tasks when allowed to use AI tools, even though they believed it had sped them up by 20%. The same Stack Overflow survey found 46% of developers distrust the accuracy of AI output, and 66% say their top frustration is AI solutions that are "almost right, but not quite."
So AI has made a custom build realistic for problems that used to be too small to justify one. It hasn't removed the need for someone who knows what good looks like to scope, check and maintain the work.
How Promptwise scopes build vs buy
At Promptwise, I scope the whole business first, across sales, marketing, operations and service, billed by the hour and agreed upfront: what you pay for, where the workarounds are, and which problems are better bought, connected or built.
Where a build makes sense, I introduce you to one of my trusted AI or build contacts with a clear brief, and in some cases I set it up myself. Start with a small prototype so you can test the idea before committing to the full build. The builder then does their own technical scoping from my brief, and quotes and contracts with you directly. See custom builds for how it works, and building custom software in the age of AI for the bigger picture.
A real example: this website
I made this call myself when I set up Promptwise. The obvious move was to buy: a website builder, a directory template for the tool pages and a form tool for enquiries. Flow Ninja's 2026 breakdown of Webflow pricing puts its Premium plan, the tier a content-heavy site needs, at US$25 a month billed annually, and that's before you pay anyone to design the site or load more than 200 tool pages by hand.
Instead, I built it with Claude. In about a week of working sessions, we produced:
- the core site, covering our services, niches, tech and results
- a tool directory with its own page for each of more than 200 tools, all generated from one editable data source
- four free calculators and an AI search visibility checker in Tools & Templates
- this Education Centre, including my Sales Philosophy
- lead forms that email me the moment someone enquires, plus the structured data, sitemap and llms.txt file that help search engines and AI assistants read the site
The site runs on Netlify's free plan, the forms use Netlify's built-in form handling with a free Google Apps Script alert, and the build used an AI subscription I already pay for. Beyond the domain name, the running cost is close to zero.
Why build rather than buy? Mainly the tool directory. I wanted every tool on its own page, structured the way search engines and AI assistants read, and updated from one source instead of edited page by page. No template did that without heavy workarounds, which is exactly the build test in the checklist above.
The trade-off is the one this article warns about: I own the maintenance. There's no support line if a page breaks, and changes go through Claude rather than a drag-and-drop editor. For a business that edits its site every day and has no appetite for that, buying would be the better call.
The short version
Buy for the standard stuff. Connect your tools to close the gaps. Build where it gives you an edge, and prove it with a small prototype first. If you're not sure which bucket a problem falls into, book a free 30-minute consult and we'll work through it together.
Sources
- 2026 SaaS Management Index findings on licence use and app counts, Zylo, 2026
- Retool's 2026 Build vs. Buy Report Reveals 35% of Enterprises Have Already Replaced SaaS With Custom Software, Business Wire / Retool, 2026
- Stack Overflow Developer Survey 2025: AI, Stack Overflow, 2025
- Anthropic Economic Index: AI's impact on software development, Anthropic, 2025
- Measuring the Impact of Early-2025 AI on Experienced Open-Source Developer Productivity, METR, 2025
- Webflow Pricing in 2026: A Practitioner Breakdown of Plans, Seats, and Add-Ons, Flow Ninja, 2026
Adam Agresta is the founder of Promptwise Advisory in Sydney. He has spent 15+ years in sales, marketing and partnerships, including running ActiveCampaign’s ANZ partner channel and building an AI start-up’s sales function from the ground up. More about Adam.