2024 was a great year for AI with lots of innovation and tech giants pouring billions into AI development, but most companies face a more practical challenge: how to implement AI effectively.
The latest Bain survey reveals a surprising truth: companies aren't struggling with AI because of technology - they're struggling with implementation. Should they build custom solutions or buy off-the-shelf? Centralize or decentralize decision-making? Let's look at what leading companies do differently.
First, AI is working
47% say the technology met expectations
43% report better business outcomes than expected
10% saw technology helped for broader change
Where Companies Are Finding Success
Most active areas are:
Software development (41% piloting, 25% in production)
Core product enhancement (49% piloting, 16% in production)
Natural language interfaces (43% piloting, 20% in production)
Two Key Differences For How Leaders Implement
There seem to be two things that set successful companies apart:
Balanced Approach
Leaders balance Do-it-Yourself (DIY) (53%) and off-the-shelf solutions (47%)
Laggards heavily rely on DIY solutions (69%)
Leading companies maintain a balanced approach, keeping significant in-house AI development capabilities while pragmatically using ready-made solutions where appropriate.
Centralized Decision-Making
Leaders maintain tighter control around picking and prioritizing AI projects while giving teams flexibility in execution.
Most Common Implementation Hurdles
Quality & accuracy concerns (44%)
Lack of expertise (44%)
Company data not ready (33%)
Unproven ROI (29%)
Why This Matters
This study busts some common myths about AI implementation. You don't need to build everything in-house, buy every new tool, or give teams complete freedom.
What works is balance: successful companies maintain a near-even split between building and buying (53-47%), keep strategic control while empowering teams to execute, and focus on fundamentals before rushing to implement.
The key isn't choosing between build or buy, control or autonomy - it's finding the right mix given your organization’s capabilities and goals.
Your Next Move
Audit your AI portfolio - are you defaulting to all-build or all-buy? Consider where proven solutions could accelerate progress versus where custom development truly adds value.
Review your governance - are key decisions centralized while giving teams room to execute? Start small, learn fast, and adjust as you go.
Those are my Thoughts From the DataFront
Max
Notable Substackers To Check Out:
The Technomist: Where business, economics, and technology meet
New Economies: Tech trends and startup ecosystem insights
AI Supremacy: Decoding AI's impact on business and society
AI Disruption: Daily insights from an AI engineering veteran
Money Machine Newsletter: Market-beating insights in 5 minutes
VC Corner: Weekly startup ecosystem analysis