Questions to Ask Before Signing with an Automation Consultant

You've decided your business needs automation help. You've talked to a few consultants, sat through their demos, and one of them sent a contract. It's sitting in your inbox right now.

Before you sign it, read this. Because 30 to 50% of automation projects fail according to Ernst & Young, and most of those failures don't start with bad technology. They start with bad agreements.

Why the Contract Stage Is Where Projects Go Wrong

The automation consulting market is booming. The global RPA market alone hit $3.94 billion in 2023 and is projected to reach $37.87 billion by 2032 (Grand View Research). That growth has attracted both excellent firms and a wave of opportunists.

A 2024 ABBYY survey of 400 executives found that 38% of automation projects fail because they are too complex, and 3 in 10 fail because the processes being automated weren't properly understood before work began. Both of those problems should be caught during the discovery and scoping phase, which means they should be caught before you sign a contract.

Here's the pattern we see repeatedly: a business owner gets excited about automation, signs a contract that's light on specifics, and six months later has spent $40,000 on a system that doesn't quite work and a consultant who's moved on to the next client. The questions below are designed to prevent that.

Questions About Their Process and Experience

These are the ones that tell you whether you're hiring a real practitioner or someone who watched a YouTube course about AI automation agencies.

"Walk me through your discovery process for a business like ours."

A good consultant will describe a structured methodology: process documentation, stakeholder interviews, workflow mapping, identifying bottleneck points, and evaluating automation feasibility for each process. If their answer is "we'll figure it out as we go" or they skip straight to tool recommendations, that's a problem.

"How many projects have you completed for businesses our size and industry?"

Automating workflows for a 10-person accounting firm is fundamentally different from automating a 200-person logistics company. Ask for specific numbers, not generalizations. If they've done 50 projects but none in your industry, their learning curve is on your dime.

"What's your project failure rate, and what caused those failures?"

Any honest consultant has had projects go sideways. The ones worth hiring will tell you about them and what they learned. If someone claims a 100% success rate across dozens of projects, they're either lying or defining success very loosely.

"Can we speak to three clients who had a similar scope?"

References should be specific to your type of engagement, not just "here's a client who liked us." Ask those references: did the project come in on time and on budget? What broke after launch? How responsive was the consultant when things went wrong?

When you're evaluating a consulting contract, the fine print matters as much as the scope of work. ContractCraft reviews contracts and flags risks automatically for $199, which is worth the investment before signing a five-figure agreement.

Questions About Pricing and Contract Structure

Automation consultants use several pricing models, and each one shifts risk differently between you and the consultant.

"What's included in this quote, and what's not?"

This is the question that catches 80% of budget surprises. A consulting quote for "$25,000 to automate your customer intake process" might cover only the build. Discovery, testing, training, documentation, and post-launch support might all be extra. According to Prioxis, licensing fees represent only 25 to 30% of total RPA project cost. The other 70 to 75% is implementation, maintenance, infrastructure, and training. Get a line-item breakdown.

"What happens if the project takes longer than estimated?"

On fixed-price contracts, the consultant eats the overrun. On hourly contracts, you do. Either way, get a cap or a not-to-exceed clause in writing. Without one, a project scoped at $20,000 can quietly become $35,000. This isn't hypothetical: PMI's 2024 Pulse of the Profession report found that 73% of projects exceed their original budget, with scope creep responsible for 52% of those overruns. IT projects specifically run an average of 45% over budget while delivering 56% less value than predicted (Standish Group).

"What are the ongoing costs after the project is complete?"

Automations require maintenance. APIs change, systems update, edge cases emerge. Budget another $1,000 to $5,000 per year minimum for maintenance on any non-trivial automation. And there are hidden cost drivers most consultants don't mention upfront: data cleanup for legacy systems, integration complexity (each connected system adds ongoing maintenance burden), and exception handling for cases the automation can't process automatically. Ernst & Young reported in December 2024 that 83% of IT leaders cite poor data infrastructure as a key factor slowing AI automation. If your data isn't clean, expect your timeline and budget to grow.

If the consultant doesn't mention post-launch costs, they're either inexperienced or hoping you'll pay them to fix things later at full rate.

The Contract Clauses That Cost Small Businesses the Most

Beyond pricing, there are contract provisions that can quietly cost you far more than the project fee itself.

IP ownership. Many consulting contracts default to the consultant retaining ownership of the code and automations they build, granting you a license to use them. That means if you switch providers, you might not be able to take "your" automations with you. Negotiate for full IP assignment of all custom work built specifically for your business. The consultant's pre-existing tools and frameworks are typically theirs to keep, and that's reasonable.

Data rights. Who owns the data your automations process? Can the consultant use your business data for training models or improving their tools for other clients? If the contract mentions "aggregated data usage" or "anonymized training data," that's your customer information being repurposed. Get clear language: your data stays yours, period.

Vendor lock-in. Some consultants build on proprietary platforms or custom frameworks that only they can maintain. If you leave, you have to rebuild from scratch. Ask explicitly: "If we end this engagement, can another developer pick up where you left off?" If the answer involves words like "proprietary methodology" or "our custom platform," proceed carefully.

Automatic renewals. Retainer contracts often auto-renew unless you cancel within a narrow window (sometimes 30 days). Set a calendar reminder, or better, negotiate for manual renewal with a 90-day notice period.

Termination clauses. What happens if you need to end the engagement early? Some contracts include penalties of 50% or more of the remaining contract value. A reasonable clause lets either party terminate with 30 to 60 days' notice, with payment for work completed.

If your automation project involves handling customer data in a regulated industry, ComplianceAI monitors compliance requirements and flags risks automatically for $249. It's particularly useful during vendor evaluation when you need to verify a consultant's compliance claims against actual regulatory requirements.

Red Flags That Should Kill the Deal

Not every red flag means the consultant is bad. But these should at minimum pause the conversation.

They guarantee specific ROI before analyzing your processes. "We'll save you $100,000 in the first year" is a sales pitch, not a professional assessment. A credible consultant will tell you ROI depends on what they find during discovery and give you a range based on similar projects.

They propose solutions before discovery. If the consultant already knows exactly what tools to use and how to set them up before spending time understanding your workflows, they're selling a templated product, not a consulting service. That's fine if it fits your needs, but you're paying consulting rates for a productized service.

They can't explain their approach in plain language. Technical jargon is a deflection tool. A good consultant should be able to explain their approach clearly enough that a non-technical business owner understands what will happen, in what order, and why.

They have no post-implementation support plan. Building an automation and walking away is like installing plumbing and leaving before testing for leaks. Automations break. Ask what support looks like for the first 90 days and beyond. If the answer is "you can hire us again at our hourly rate," factor that cost into your decision.

Their background doesn't add up. The AI boom has created a wave of people rebranding from unrelated fields. Over 80% of "AI Automation Agency" content on YouTube and social media is clickbait from people with no real implementation experience (DowneLink, 2025). Check their LinkedIn history. If they were selling crypto courses or dropshipping masterclasses 18 months ago and are now "AI automation experts," be skeptical. For legitimate consultants, look for platform-specific certifications like UiPath Certified Developer, Automation Anywhere Certified Professional, or Microsoft Power Automate credentials, and ask for measurable outcomes from past projects, not just logos on a website.

When You Don't Need a Consultant at All

This is the part most consulting articles won't include, but it's important.

If your automation needs are simple, you probably don't need a consultant. Standard workflows like email sequences, appointment reminders, CRM data entry, invoice processing, social media scheduling, and basic customer support triage can all be handled by off-the-shelf tools.

Zapier starts at $19.99/month. Make starts at $9/month. Purpose-built AI tools handle specific functions for $99 to $299.

Here's a simple test: describe the automation you need in one sentence. If you can ("when a form is submitted, add the contact to my CRM and send a welcome email"), you can set that up yourself in an afternoon with a no-code tool. If you can't describe it simply, or if it involves multiple systems talking to each other in complex ways, a consultant might be worth the investment.

The breakpoint tends to be around $2,000 to $5,000 per month in manual labor costs. Below that, the math rarely justifies consulting fees. Above it, the ROI timeline shortens enough to make sense. McKinsey estimates a 200% ROI in the first year for well-implemented automation, but "well-implemented" is doing a lot of work in that sentence.

Our advice: start with the specific function that costs you the most time, try an off-the-shelf solution first, and bring in a consultant only for the problems that off-the-shelf tools can't solve. You'll make a better-informed hire because you'll already know what your automation needs to do.

Need to automate specific business functions without hiring a consultant? Automiva builds AI tools for lead scoring, customer support, accounting, compliance, and more. Start with the function that hurts the most.