Your business runs on 87 apps. Half the licences are idle.

The average small business now runs somewhere between 87 and 152 SaaS applications. When people first hear that they assume it's wrong. Then they open their card statement and start counting.

It adds up quietly, and always for a good reason at the time. A scheduling tool. A separate invoicing tool because the scheduler's invoicing was weak. A forms tool because the website's forms didn't do conditional logic. A spreadsheet that quietly became the actual database. Then AI tools — the average small business went from 0.8 AI subscriptions in 2023 to 3.2 now.

The waste is measurable: roughly 53% of SaaS licences sit idle, and companies waste 17–25% of their software budget on tools nobody meaningfully uses.

The cost that isn't on the invoice

Cancelling dead subscriptions is the easy, obvious win, and it's not the interesting one.

The real cost is that none of these tools talk to each other, so a person does it. Someone copies the booking into the calendar. Someone re-types the enquiry into the CRM. Someone exports a CSV on Friday, pivots it, and pastes the numbers into a WhatsApp group.

That work is invisible because it never appears as a line item. It shows up as your operations person being busy, as things taking three days, as the lead nobody followed up. You are already paying for integration — in salary, in delay, and in the deals that leaked out of the gaps.

Software sprawl doesn't feel like a software problem. It feels like needing to hire someone.

Two different fixes, and they're often confused

Automation is for work with no conversation in it. A trigger happens, conditions get checked, actions fire. Form submitted, so create the CRM record, send the acknowledgement, notify the right person by territory, add the follow-up task for three days out. Nobody starts it and nobody watches it. That's business automation.

An AI agent is for work that requires judgement or dialogue. Answering a customer who phrased their question in a way you didn't anticipate, qualifying a lead by asking the right follow-up, drafting a reply that needs to sound like a person. That's an agent, and it's a different tool with a different price.

The mistake we see most often is paying for AI where a trigger would do. An LLM to move data between two systems is slower, more expensive, and less reliable than the boring automation that has existed for a decade. Use the cheap thing for the deterministic work.

The opposite mistake is rarer but more damaging: trying to script your way through something that genuinely needs judgement, ending up with forty branching rules that nobody dares to touch.

What we'd actually do first

Not "automate everything." Start smaller and more honest:

1. List every recurring task that a person starts. Not what they do — what makes them start doing it.
2. Mark the ones with no decision in them. Those are automation candidates today.
3. Count how many tools each one touches. Anything crossing three or more systems is where the delay lives.
4. Automate the highest-frequency one only. Then measure for a fortnight before adding the next.

Most businesses find their first genuine win in lead routing and follow-up, because that's where the delay costs actual revenue rather than just patience.

The version where you buy less software

The other route out is fewer tools, not better glue. That's the thinking behind Lilium Studio — storefront, CMS, bookings, intake, invoicing and payments in one place, so there's nothing to integrate because it was never separate.

That won't suit everyone. If you're deeply committed to a stack that works, replacing six tools is a bigger disruption than connecting them, and we'll say so. But if you're early enough that the sprawl hasn't set, buying one thing instead of nine is the cheaper decade.

Either way the first step is the same and it's free: count the tools, and count how many times a day a human moves data between them. The second number is the one that should worry you.

---

Sources: SaaS tool sprawl statistics 2026, Breeze · Zylo, 70+ SaaS statistics · 2026 SaaS waste report for SMBs