When a company decides to put AI into its operations, the first practical question is who will build it. The three real options are an AI automation agency, an AI consultant or consulting firm, or your own engineers. They are not interchangeable.
AI automation agency
Good at: repeatable automations on common tools, such as connecting your CRM, forms and email with AI steps. Watch for: templated solutions that fit the average client, and handing off to juniors after the sales call. Ask who actually builds your project and who owns the result.
AI consultant or consulting firm
Good at: deciding what to build and why, scoping, architecture, vendor and model selection, and often building the first production version. Watch for: strategy decks with no implementation path. A good partner scopes, builds and hands over, or stays honest about where their work ends.
In-house team
Good at: deep knowledge of your domain, and long-term ownership. Watch for: the time to hire and ramp, and the fact that AI engineering is a different skill set from general software engineering. Without evaluation practices, in-house prototypes often stall before production.
Side by side
Outside partner
- Starts in days or weeks
- Brings patterns from similar builds
- Fixed scope and price possible
- You must plan the hand-over
In-house
- Months to hire and ramp
- Deep domain knowledge
- Ongoing salary cost
- Owns it continuously
A hybrid that works for many teams
Use an outside partner to build the first production system quickly, with a requirement that code, prompts, evaluation suite and runbook are handed over. Your engineers then own and extend it, with a monitoring retainer if you want one. You get speed now and ownership later.
Questions to ask any AI partner
- Who will do the work, and can I speak to them?
- Do I own the code, prompts and data when it is done?
- How do you measure that it works, on my cases?
- What does it cost, in total, before I commit?
- What happens after launch?
More on vetting vendors in How to choose an AI consulting firm, and how we work in What an AI consulting engagement looks like.
Frequently asked questions
What is the difference between an AI agency and an AI consultant?
Agencies typically deliver packaged automations on common tools. Consultants focus on deciding what to build, architecture and often the first production build. The line blurs, so ask who does the work and who owns the result.
Is it cheaper to build AI in-house?
Over a long horizon, sometimes. In the short term, hiring and ramp-up time usually make an outside partner faster, and a fixed-price build makes the cost predictable.
How do I avoid vendor lock-in with an AI partner?
Require hand-over of code, prompts, evaluation suite and documentation, deployed in your own infrastructure, and keep your data in portable formats.
Related: AI strategy sprint · AI agent development
Want a straight answer for your project?
Book a 30-minute call. I will tell you what I would do, and a fixed price if it makes sense.