Should you build AI in house or hire a partner?
A lot of owners and operators I talk to hit the same fork in the road. They know AI should be doing real work in the business. Then the question comes up: do we build this ourselves, or do we bring someone in?
I'm Andy Robbins, co-founder of AI Answered. We help established companies grow, so I'm not neutral here. But I'll give you the honest version, including when you should not hire anyone.
If you already know you want an outside set of eyes, you can talk with us about a growth partnership. If not, keep reading.
What the research actually says
In 2025, MIT's Project NANDA put out a report called The GenAI Divide: State of AI in Business 2025. The headline number got passed around everywhere: about 95% of organizations were getting zero return from generative AI.
The part that got less attention is more useful. According to Fortune's coverage of the report, buying from specialized vendors and building partnerships worked about 67% of the time, while internal builds succeeded only one third as often.
That doesn't mean your team can't build anything. It means the default move a lot of companies make, "let's have IT figure it out," fails more often than people expect.
Why in house builds stall
It's rarely because the team isn't smart. It's usually one of these:
- Nobody owns it. AI becomes a side project for someone who already has a full time job.
- It starts with a tool, not a problem. Someone gets excited about a chatbot and works backward to find a use for it. I wrote more about that in People before tools.
- No one knows the numbers. If you can't say what a process costs today, you can't say whether AI made it better. That's the whole point of AI won't fix a business that doesn't know its numbers.
- It never touches daily work. The MIT report points to brittle workflows and tools that don't fit day to day operations as a big reason projects fail.
When building it yourself makes sense
You should probably keep it in house if:
- You have someone whose actual job is this, with the time and authority to change how work gets done.
- The use case is narrow and well understood, like summarizing calls or drafting first pass documents with tools your team already has.
- You already track the numbers for the process you want to change.
- You're fine moving slower in exchange for building the skill internally.
If that's you, start with where to start with AI in your business. You may not need a partner yet.
When a partner makes sense
Bring someone in if:
- You've tried a few tools and nothing stuck.
- AI keeps getting pushed to next quarter because nobody has the bandwidth.
- The work touches revenue, operations, or how teams hand off to each other, where a mistake is expensive.
- You want someone accountable for a result, not just a deliverable.
A good partner should leave your team stronger, not dependent on them. If you're comparing options, read what to look for in an AI consultant and what an AI consultant actually does.
How we approach it at AI Answered
We look at a company's numbers first. Where is time going, where is money leaking, what would actually move the business. Then we price on the value we can deliver, so you're never staring at a vague number wondering what you're paying for.
Sometimes the answer is that your team can handle it, and we'll say that. Sometimes the problem isn't AI at all. That's fine too.
A simple way to decide
Ask yourself three questions:
- Who on my team owns this, and is it their actual job?
- Do I know what this process costs today?
- What happens if this takes another year to figure out?
If you can't answer the first two, or the third one makes you uncomfortable, it's probably time to bring someone in.
If you want to see how we'd look at your business, start a conversation with AI Answered. And if you're the one who has to explain AI to people who didn't grow up with it, our book AI for Boomers keeps it simple.
Sources: MIT Project NANDA, The GenAI Divide: State of AI in Business 2025 (linked above); Fortune, "MIT report: 95% of generative AI pilots at companies are failing," August 18, 2025 (linked above).