Nine tools is a short list. Where’s everything else?
Not in production here, so not on the page. We could rank forty products from their marketing sites in an afternoon and it would be worth nothing to you. The short list is the honest one.
Nine systems, not ninety
Nine tools, because that’s how many we’ve actually used.
Most "best automation tools" lists rank forty products the writer has never run. This is the nine systems Kindra has actually built against, what each is genuinely for, and where each stops being the right answer — including the two automation platforms we use ourselves and the CRM we run our own business on.
What we actually run
Nine systems. That is the complete list, and it is short on purpose — a tool we have not put into production is a tool we cannot tell you anything useful about.
Three accounting packages, two job management systems, one payments platform, one CRM, and two automation platforms. We also run one of these on our own business rather than only for clients, which is worth saying because it changes how much you should trust the opinion.
- Xero, MYOB, QuickBooks Online Accounting. All three implemented against. These are where the money actually lives, so they are almost always one end of any integration we build.
- ServiceM8, Simpro Job management, at opposite ends of the market — <a href="/compare/servicem8-vs-simpro">the difference is structural rather than featural</a>.
- Stripe Payments. In production. The one on this list with the best-documented developer experience, which matters more than it sounds.
- GoHighLevel CRM and marketing. We run Kindra’s own client pipeline on it, which means the opinions here are from daily use rather than an implementation.
- Zapier, n8n Automation platforms, both in production. <a href="/compare/zapier-vs-n8n">Where it runs and how it meters is the whole decision.</a>
By job: which tool for which problem
The useful question is not "what is the best tool" but "what is the smallest tool that closes this specific leak". Almost always the answer is something you already pay for, connected to something else you already pay for.
| What we use | Where it stops | |
|---|---|---|
| Moving data between two systems | Yes Zapier for low volume and visible logic; n8n where volume or complexity makes task billing hurt. | Partly Past roughly a dozen steps, or where silent failure costs money, both stop being the right shape. |
| Getting paid | Yes Stripe. The developer documentation is the best on this list and that shows up as fewer surprises. | Partly Fees are a real cost at volume. Worth modelling against bank transfer for large invoices. |
| Tracking jobs in the field | Yes ServiceM8 for small teams billing by job volume; Simpro where the work is projects rather than jobs. | Partly Neither will fix a scheduling process that changes weekly. Stabilise first. |
| Keeping the books | Yes Xero, MYOB or QuickBooks Online. We have implemented against all three and will not rank them here. | No None of them will read a supplier invoice that arrives differently every month. That is a separate build. |
| Chasing leads and clients | Yes GoHighLevel. We run our own pipeline on it, which is the only reason we are willing to comment. | Partly It is a large product and most businesses use a fraction of it. Sprawl is the risk, not capability. |
What we tried and dropped, and why
This is the section every list of this kind should have and almost none do, and we are going to be honest about why ours is thin: Kindra AI was founded in 2025, so the graveyard is short. Nine systems in production, and none of them have been abandoned.
What we have dropped is narrower — specific approaches rather than whole products. Chief among them: reaching for an AI step where a rule would do. It is the most expensive habit in this field, it is easy to fall into because the AI step is more interesting to build, and it costs the client every month in model usage for work a conditional would have done for free.
When something does get dropped, it will be named here with the reason. A list that only ever grows is a marketing asset.
Cost at small vs real volume
Every tool on this list is cheap at the volume you are at when you buy it. The question that decides whether you resent it in two years is what the meter counts.
Per-seat products punish hiring. Per-task products punish success. Per-job products — ServiceM8 is the example on this list — punish growth in the specific way that at least matches how you grow. Self-hosted products punish nothing and charge you attention instead, which is why where a thing runs is a commercial question as much as a technical one.
Model it at the volume you expect in three years, not today. That is the single most useful hour in any software decision and almost nobody spends it.
Where the free tiers stop working
Free tiers are genuine and worth using — ServiceM8 publishes a free tier at 30 jobs a month aimed at sole traders, and n8n’s self-hosted Community edition is free indefinitely with, in their words, "almost the complete feature set".
They stop working at predictable points: when you need a second person in the system, when the volume cap starts blocking real work, or when you need something the paid tier gates like SSO. None of those are traps — they are just where the product stops being free.
The one genuine trap is API access sitting behind a higher tier. If you plan to automate on top of a product, check that before you buy, because every future integration then carries that tier as a hidden prerequisite. It is on the checklist in how to choose job management software for exactly that reason.
When no tool fits and you need something built
Rarely, and later than most people conclude. The honest sequence is: buy the product, connect it to the other product, and only then consider building — which is the argument on custom vs off-the-shelf.
The exception is the step where a human reads something unstructured. No tool on this list will read a supplier invoice that arrives in a different format every month. That is the one place we reach for an AI step without arguing, and it is a narrow place.
If you want the framework rather than the list, the build-vs-buy decision takes about an hour and usually returns "buy, then automate the joins".
Objections
You run GoHighLevel yourself — isn’t that biased?
Yes, and that is why we said so rather than presenting it as neutral analysis. Daily use makes an opinion better informed and more partial at the same time. Weigh it accordingly.
Book a free Leak Check
20 minutes. We find where the money's going. No pitch.
Book a free Leak Check