Any vendor who gives you a price before understanding your operation is selling you a package, not a solution. But that doesn't mean the cost is a mystery. It can be thought through clearly if you split it into three layers.

The three layers of cost

  • Tools. Messaging, automation and AI model platforms are paid per use or per month. For a service SMB it's usually the smallest layer, and often you're already paying for tools you don't use well.
  • Design and implementation. Understanding the process, writing the criteria, building the flow, testing it with the team. This is where the value is — and where the difference shows between a flow the team uses and one it abandons.
  • Maintenance and adoption. Adjustments when the process changes, training, reviewing what goes out. It's the layer almost nobody budgets and the one that decides whether the investment lasts.

What is proportional

An investment is proportional when the problem it solves costs more than the solution. Three questions clarify it:

  • How many hours a month go into this task? Hours of people who could be doing billable work.
  • What does the error or delay cost? A client who didn't get a reply, a collection that came late, a forgotten agreement.
  • What would happen if the key person weren't there? The cost of dependence rarely shows up on an invoice, but it's the biggest one.

If the answers are small, the solution should be small: a template, a training session, a well-designed sheet. If they're big, a flow, an assistant or a system is justified.

What isn't worth paying for

  • A big system before having a clear process. It automates the disorder and is expensive to fix.
  • “All-in-one” tools nobody on the team will use. The license is the least of it; the real cost is the adoption that never happens.
  • Automating what happens once a month. The design costs the same as for daily work and the return doesn't exist.
  • A chatbot that replies on its own in a business where trust is the product. What it saves in replies it loses in relationship.

How to ask for a proposal you can actually compare

Ask it to separate the three layers, say which concrete process it solves, who reviews before anything goes out and what happens when the process changes. A proposal that doesn't answer that is selling technology, not capacity.

The right cost of automating is the one that's smaller than the problem — and the problem can almost always be measured in hours and in dependence.

The first step

Before asking for prices, measure: how many hours a month go into the task you want to solve and what happens when it goes wrong. With those two numbers, any proposal can be evaluated in five minutes.

MINI-DIAGNOSIS

Where is capacity being lost?

Answer five questions and detect whether the bottleneck is in follow-up, knowledge, key people or repetitive work.