Skip to content

How to test a provider with a pilot

Test a provider by asking for a free proof of concept, then buying one small paid pilot that runs on your real data while you watch, before you commit to the full build. You need a written proposal and a working piece, not a consulting engagement or a slide deck.

Owners who once paid for a project that never worked ask the same thing: how do I find out whether this provider can deliver before I hand over the full budget? This page answers that, and covers whether you need a consultant before anything is built.

Do you need automation consulting before you build?

Section titled “Do you need automation consulting before you build?”

Usually not. Most owners need one conversation and one written proposal, not a paid consulting engagement. The conversation turns the problem into something specific. The proposal shows you the system, the tools it touches, the fixed number and the date. That is what most people are really buying when they buy consulting.

Consulting on its own is the right call when the question is about the organisation rather than a system: who should own a process, what the business should stop doing, how a team should be structured. That is not our work, and we say so in the first reply. If what you need is for the follow ups to happen, skip the engagement and ask for a proposal.

The middle case is an owner who cannot yet name the problem. The discovery call exists for that. We ask questions until the problem has a shape, and it costs nothing.

What should the proposal say before any pilot starts?

Section titled “What should the proposal say before any pilot starts?”

Five things, in words you can read without a technical background:

  • the problem, written in your words
  • the system the provider will build
  • every tool it reads from and writes to
  • one fixed number
  • one delivery date

Then check what it leaves out. A good proposal names what will not be touched. A document that lists only what it will do is hiding its own scope, and the argument comes later. If a proposal could have been sent to any business, nobody has looked at yours yet.

Free first, then small and paid, then full. Each step answers one question before you spend more.

StepWhat you getWhat it settles
Free proof of concepta proof of concept or minimum viable product, free of chargethe project can be delivered at all
Small paid pilotone working piece on one channel, on your real datait works on your records, not a demo set
Full buildthe whole system, tested with you watchingit does the full job, and it is handed over to you

You can go from the proof of concept straight to the full build if you are already confident. The pilot is there for owners who want one more check before the larger decision.

Send the enquiry form and message us on WhatsApp. We set up a discovery call as soon as possible.

Four things. If any one fails, stop.

QuestionA pass looks like
Did they understand the problem?the proposal names your tools, channels and process
Does it work on real data?it runs on your leads, orders or bookings while you watch
Can you read the result yourself?you can tell it worked without a call to explain it
What do you keep?the working piece, and a clear view of the full scope

The last row is easy to forget. A pilot that leaves you with nothing but a slide deck is a demo you paid for.

The one that already costs you money quietly. It is almost always a follow up that does not happen, rather than a process that is slow.

What the owner describesA first system worth piloting
Leads arrive and nobody replies fast enoughlead outreach and follow up, see lead generation systems
Renewals lapse because nobody chased thema renewal reminder inside the CRM, see CRM and sales automation
Staff copy details between two screensa connection between those two tools, see integration
The month end report takes days by handreporting built from records you already keep

The first row is where our clearest result came from; it is written up on lead generation systems. It was not an ambitious build. It was the smallest system that closed the most obvious gap, which is what a first project should be.

What are the warning signs during a pilot?

Section titled “What are the warning signs during a pilot?”

Three signs are enough to walk away:

  1. No small option. A provider sure of their work will sell you the smallest version, because they expect the second project.
  2. A price before the questions. A number that arrives before anyone asked about your tools is a guess.
  3. A demo on invented data. A polished screen full of made-up records shows the design, not whether it copes with your real, messy data.

Be wary, too, of any provider who promises a business result, or a top spot on Google when Google itself says no one can guarantee a #1 ranking. Check their reviews the same way: the FTC has a rule on fake, false or deceptive consumer reviews and testimonials. We never sell a business result. We show what the system does on your data and let you judge it.

A pilot proves a system can handle enquiries. Something has to bring those enquiries first, and for most small businesses that is the website. If yours is hard to find on Google (not every page Google processes ends up in its index) or has no clear way to enquire, a website plan may be the better first test.

Websites are the one thing we price in public, and we build them first: you pay within 3 days of going live, with unlimited revisions until you are satisfied with the design. The 5-page plan is $97 a month, and the plan for 10 pages and above, with SEO, GEO and AEO, is $297 a month, billed yearly in US dollars. See pricing.

Tell us what went wrong last time, or what keeps slipping now. We ask questions, then send a written proposal with the free proof of concept and the pilot described in it. The full steps are on how we work, and how a custom build is priced is in how a custom AI agent is quoted.

Send the enquiry form and message us on WhatsApp. We set up a discovery call as soon as possible.

By Raze, Founder. Updated .