If you run a small service business, you’ve probably been pitched automation more than once this year. Most of those pitches start with a tool: a chatbot, an AI receptionist, a new CRM. This post explains what an AI automation consultant should do before any of that, what the work costs, and how to tell a consultant from someone selling one product.
What an AI automation consultant does
The job starts with finding where time and money leak out of the business. That means looking at how a lead comes in, who answers it, how a quote goes out, how a job gets scheduled, how an invoice gets paid, and where each of those steps stalls.
A good consultant maps the work the way it actually runs, which is rarely how the owner describes it. The person who answers the phone often has a sticky-note system nobody else knows about. Follow-ups live in someone’s head. Two tools hold different versions of the same customer.
From that map, the consultant decides what to fix first. Usually it’s the step where the business loses the most money for the least effort to fix. Just as often, the answer for a given step is “don’t automate this yet,” because the process underneath is broken and automating it would only make the mistakes faster.
Then comes the build: connecting the tools, setting up the assistant or the follow-up sequence, testing it with real cases, and handing it over with notes your team can follow. The last part is checking that it keeps working. Automations break quietly when a form changes or a password expires, so someone has to watch them.
How that differs from an AI automation agency or a freelancer
Many agencies and freelancers start from a product. They build chatbots, so the answer is a chatbot. That isn’t dishonest, and if you already know exactly what you need, it can be the fastest way to get it. The risk is buying a tool for a problem you don’t have.
Price is the other difference. A freelancer selling one tool quotes for that tool. A consultant’s price depends on what the diagnosis finds, which is why the same title can mean a $1,000 project or a $10,000 one.
When the answer isn’t automation
A roofing contractor came to us wanting an AI chatbot, an AI receptionist and a roof calculator for his website. Once we looked at the numbers, response time wasn’t the problem. Not enough leads were reaching the site to begin with, and a faster reply to a lead that never shows up doesn’t help.
So the project changed. Instead of adding more tools to the website, we built a way to find homeowners likely to need a roof, using building permits and storm history. It wasn’t what he asked for on day one, but it went after the problem he actually had.
What an AI automation consultant costs
There are two ways in, and they cost different amounts.
If you already know which tool you need. Sometimes the answer is clear before any audit. You know you need an assistant that answers website and WhatsApp messages, or the Free Workflow Review makes it obvious in 30 minutes. In that case we skip the audit and go straight to setup. These projects usually run $1,000 to $5,000 to implement.
If you need the workflow figured out first. Most of our work starts with a Workflow Audit: Focused (one week) from $1,500, or Full (two weeks) from $4,500. You get a map of how the work runs today and a build order: what to fix first, what it should cost, and what to leave alone. Builds after an audit usually land between $2,000 and $10,000, and we set the exact price once the audit is done. If you sign within 90 days, the audit fee is credited toward the build.
Why the range is so wide
The tool matters more than anything else. Two examples from either end:
- An AI assistant for your website or WhatsApp. Most of the work is training the assistant on your services, prices and policies, testing it with real questions, and deciding when it hands a conversation to a person. That sits near the bottom of the range, around $1,000 to set up. After that, the ongoing cost depends on volume: the assistant runs on usage credits, so a business with a few dozen conversations a month pays far less than one with thousands.
- A full automated workflow. Think of an event planner’s whole booking process: the inquiry comes in, gets logged in a CRM like ActiveCampaign or GoHighLevel, triggers a proposal, sends a contract, collects a deposit, and starts a follow-up sequence, with Zapier or a similar tool connecting each step. Building it is half the job. The other half is making sure it doesn’t break, and when it does, having logs that show which step failed so it can be fixed quickly. That sits at the top of the range.
Other things that move the price:
- How many workflows are involved (lead response alone, or lead response plus follow-up plus reporting)
- How many tools need to talk to each other, and whether they have usable integrations
- How messy the data is, since duplicates and missing fields have to be cleaned before anything is automated
- How many people use the system and need to be trained on it
- Languages, if customers need to be answered in both English and Spanish
After launch, some clients pay a monthly fee for monitoring and changes, and some take it over themselves. Either way, get it in writing before you start.
Red flags when hiring one
- A firm price before anyone has seen your process. A typical range is useful, and it’s fair to publish one. A fixed quote for your build before anyone has looked at how your business runs is a guess.
- Revenue promises. No one can honestly promise that automation will add a specific dollar amount to your revenue. Time saved can be estimated. Revenue depends on too many things outside the build.
- The same tool for every client. If every case study ends with the same product, that product is what’s being sold.
- No plan for after launch. If nobody says who watches the automations once they’re live, the answer is nobody.
Questions to ask before you hire
What will you look at first? You want to hear about your workflow and your numbers, not about a platform.
What would make you tell me not to automate this? A consultant who has never said no to a project hasn’t been diagnosing.
Who owns the accounts and the data when we’re done? The tools should be in your name, with your logins. If the consultant leaves, your systems should keep running.
How will we know it’s working? Agree on one or two numbers before the build starts, such as how fast new leads get a reply (more on that in what speed to lead is) or how many hours a week go to manual follow-up, and check them after launch.
