You have decided your revenue team needs AI that does real work. It researches accounts, updates the CRM, routes leads and drafts the follow-up. Now you need someone to build it.
Search for help and you land in two different worlds. "AI consulting firms" returns the strategy houses and long lists of the top fifty. "AI automation agency" returns workflow builders, plus a page of videos on how to start one.
Either can be the right call. They sell different things on different timelines, and you end up owning different amounts of what gets built.
This guide compares four options for a revenue team at a company of 100 to 300 people. Magnetiz builds revenue agents for a living, which puts us in the specialist group, so weigh our view with that in mind.
What are your options for building AI into a revenue team?
Four options, sorted by who does the work and who keeps it.
- Large consulting firms. Accenture, Deloitte, IBM Consulting, McKinsey's QuantumBlack and BCG X, selling strategy, transformation programs and, more and more, their own platforms.
- Specialist AI consultants. Smaller firms and independents who design and build for one function or industry.
- AI automation agencies. Teams that build workflows on tools like n8n, Make and Zapier, usually for a setup fee and a monthly retainer.
- In-house. You hire an AI or GTM engineer and build it yourself.
The labels blur at the edges. Plenty of specialists build workflows, and some agencies do strategy. In practice the options separate on three things.
The size of the engagement, how long it takes to reach a working system, and who runs it after the builders leave.
Watch for a proposal that fits one label on the sales call and another in the contract.
What do large AI consulting firms deliver?
Scale, strategy and a growing platform business.
The large firms have built AI practices at a scale nobody else matches.
Accenture reported $5.9 billion in generative AI new bookings for its 2025 fiscal year, the last full year it broke the number out.
BCG X describes itself as BCG's tech build and design unit, with more than 3,000 tech builders.
Their offer is moving from people toward products.
In March 2025 Deloitte launched Zora AI, a set of agents sold in a cloud subscription model and customized for each client.
When a firm's platform sits under the system it builds, part of what you buy is a subscription you keep paying.
These firms fit best when AI is one piece of a larger change, such as a merger, a new operating model or a regulated rollout across several countries.
None of them publishes engagement minimums, so ask early what size of project the team in front of you usually takes on.
Watch for a strategy deck that ends with a recommendation to buy the firm's own platform.
What does a specialist AI consultant do?
A smaller team that designs and builds for one function.
Specialist consultants range from independents to firms of a few dozen people. The strongest know one function deeply, such as revenue operations, finance or support, and build inside the tools that function already runs.
No independent survey tracks what this group charges, and the guides that quote hourly rates are written by consultancies selling the service.
Ask instead for a fixed scope tied to one working system, so every quote you collect prices the same thing.
The trade is depth for breadth. A specialist can get a revenue agent running in your CRM in weeks. Running your company's wider transformation is a different job.
Watch for a specialist who can describe the build and cannot show one running today.
What does an AI automation agency build?
Workflows that connect your tools, priced as setup plus retainer.
An AI automation agency builds the connections between tools you already pay for. A form fill becomes an enriched lead, a routed record and a drafted reply. The common platforms are n8n, Make and Zapier.
Pricing usually comes in two parts. One guide to starting an agency puts setup fees at $2,000 to $50,000, with monthly retainers from $500 to $3,000 or more.
It is written for people starting agencies, so treat the range as a sketch.
The category is young and growing fast. Search "AI automation agency" today and page one includes videos and a course community on how to start one.
The experience behind the label runs from a few months to many years, so ask how long the team has been running systems for clients.
The question that matters most is where the workflows live. When they run in the agency's own accounts, the system leaves with the retainer.
Watch for workflows built in an account with the agency's name on it.
When does building AI in-house make sense?
When you can hire someone who will stay to run it.
Building in-house gives you full control and the slowest start. You need someone who can design agents, connect them to your systems and keep them running.
That person is expensive.
Levels.fyi puts median total compensation for machine learning engineers in the US at $280,000, a figure weighted toward large tech companies.
The closest federal category, computer and information research scientists, had a median wage of $140,300 in May 2025.
In-house works best once you know what the first system has to do. If you are still deciding, a first system built with outside help shows you what the hire needs to be good at.
Watch for a job description written before anyone has decided what the first system does.
Does outside help work better than building alone?
The best available data says partnerships reach production more often.
MIT's NANDA initiative interviewed and surveyed companies for its State of AI in Business 2025 report.
External partnerships reached deployment about 67% of the time, compared with about 33% for tools built internally. The figures are self-reported, and the report notes the link does not prove cause.
Two details matter when you read it. "External partnership" in the report mostly means buying from AI vendors, which is wider than hiring a consultant. And it measures reaching deployment, the first finish line in a longer race.
The same report found that top mid-market companies moved from pilot to full implementation in about 90 days on average, against nine months or longer for large enterprises.
At 100 to 300 people, speed is one of the advantages you start with.
Watch for a proposal whose timeline looks like the enterprise one.
How do the four options compare side by side?
What you get, how fast, and what stays with you.
| Large firm | Specialist consultant | Automation agency | In-house | |
|---|---|---|---|---|
| What you get | Strategy, programs and platforms | One function designed and built | Workflows between your tools | Full control |
| Time to a first working system | Months | Weeks | Days to weeks | Months, starting with the hire |
| What you own at the end | Depends on the platform terms | The build, when the contract says so | Depends on whose accounts it runs in | Everything |
| Who runs it after | The firm or its platform | Your team, after a handover | Often the agency, on retainer | Your hire |
| Best fit | Company-wide change | One function with a clear job | Connecting tools quickly | A company that knows what to build |
The timelines are typical ranges, and every engagement varies with scope and data.
Watch for a proposal that leaves the "who runs it after" row blank.
Which option fits a 100 to 300 person revenue team?
Match the option to the job, then check the ownership.
Choose a large firm when the AI work is part of a company-wide program and you want one partner across many functions.
Choose a specialist consultant when one function has a clear job for AI, such as research, routing or follow-up, and you want it built in your own systems with a handover.
Choose an automation agency when the job is mostly connecting tools you already pay for, and a retainer to keep it running suits you.
Build in-house when you already know what the first system does and can hire someone who will stay to run it.
Whichever you pick, the deciding questions are the same. Where does it run, who reviews the output, and what do you own when the engagement ends?
Our guide to the questions to ask before hiring an AI consultant or automation agency turns them into a scorecard.
What should you own whichever option you choose?
The code, the prompts, the criteria and the accounts it runs in.
Ownership lives in the contract, and small words carry it.
The law firm Morgan Lewis advises looking for language where the supplier "hereby assigns" rights, and checking whether it assigns "its" work product or "the" work product, because the two cover different things.
For a revenue system, ask for four things in writing. The code and workflows, the prompts, the criteria used to check the output, and the accounts or workspaces where it all runs.
If any of them stay with the vendor, ask what happens to the system the day you stop paying.
Watch for a contract that grants you a licence to use what you paid to have built.
The pattern across the four options
Each option can get AI working in a revenue team. They differ in how much of it is still yours, and still running, a year later.
Pick the one that leaves you a system your own people can run.
---
FAQ
What do AI consulting firms do?
AI consulting firms help companies decide where to use AI and then build or buy the systems to do it.
Large AI consulting firms such as Accenture, Deloitte, IBM Consulting, McKinsey's QuantumBlack and BCG X combine strategy with build teams and, more and more, their own platforms.
Smaller AI consulting firms focus on one function or industry and build inside the client's existing tools.
What is an AI automation agency?
An AI automation agency builds workflows that connect a company's tools and add AI steps, such as enriching a lead or drafting a reply.
Most AI automation agencies work on platforms like n8n, Make and Zapier and charge a setup fee plus a monthly retainer. The category is young, so the experience behind the label varies widely.
What is the difference between an AI consultant and an AI automation agency?
An AI consultant usually designs and builds a system for one function, scoped to a business outcome and handed over to your team.
An AI automation agency usually builds and maintains workflows between tools, often on an ongoing retainer. For either, the useful test is where the system runs and who runs it once the engagement ends.
How much does an AI automation agency or AI consultant cost?
No independent survey tracks AI consultant or AI automation agency pricing in the US. One guide for new agencies cites setup fees of $2,000 to $50,000 plus retainers of $500 to $3,000 or more a month.
Ask every firm for a fixed scope tied to one working system, so you compare quotes on the same terms.
Should you build AI in-house or hire an AI consulting firm?
Build AI in-house when you know what the first system should do and can hire someone to run it for the long term. Machine learning engineers in the US carry a median total compensation of about $280,000 on Levels.fyi.
Hire an AI consulting firm or specialist when you need the first system working sooner. MIT NANDA's 2025 report found external partnerships reached deployment about twice as often as internal builds, on self-reported data.
Which are the largest AI consulting firms?
The largest AI consulting practices sit inside Accenture, Deloitte, IBM Consulting, McKinsey (QuantumBlack) and BCG (BCG X).
Accenture reported $5.9 billion in generative AI new bookings for fiscal 2025, the last full year it reported the figure, and IBM reported a generative AI book of business above $6 billion inception to date in early 2025.
Their size suits company-wide programs, and a revenue team with one clear job may move faster with a smaller firm.