Lowe’s deepens B2B push
Lowe’s announced an $8.8 billion acquisition of Foundation Building Materials to bolster its presence in the $250 billion professional builder market. This follows Home Depot’s similar pivot toward B2B, bringing its business to approximately 45% B2B revenue. The hosts noted that B2B offers higher margins, better customer lifetime value, and more stable revenue than consumer retail.
The hosts observed that companies starting in B2C and moving to B2B never go back, while B2B companies that try B2C often return. B2B customers are more loyal and negotiate less aggressively, though the sales cycle is longer and requires more relationship development.
The 95% failure rate
MIT Media Lab research found that 95% of generative AI pilots at companies are failing to deliver measurable financial results. Only a tiny fraction of initiatives scale to production. The study identified that the real barriers are not technical but strategic, organizational, and cultural.
The genius is not in the invention. It is in the operationalizing of the invention. Thomas Edison did not invent the light bulb. His real genius was lighting New York City.
Andy Hoar, Master B2B
Projects fail because they are undefined
Many AI projects enter what MIT calls pilot purgatory, stalling because they are not designed to scale or plugged into actual systems. The hosts gave an example of using AI to optimize pricing by benchmarking existing prices after the fact. The problem is that AI should be used at the beginning of the pricing process, not the end. Without redefining the fundamental process, experiments cannot demonstrate value.
Projects fail because they are unfocused
AI projects that are too broad uncover exponential complexity. Implementing step one may reveal three other things that need changing, each of which reveals three more. A company that tries to use AI to transform how the entire finance organization works will never succeed because it will uncover dozens of interdependent areas requiring fixes.
The hosts recommended targeting specific tactical problems. Fix a backend process. Automate invoice generation. Use AI to solve a defined problem, then expand. The hosts noted they are deploying AI in their own business by focusing on one specific area rather than trying to transform everything at once.
If you bite off more than you can chew, you will choke. It is a highly iterative process. You discover things as you go along.
Andy Hoar, Master B2B
Projects fail because they are unmeasured
Without metrics defined upfront, AI projects cannot demonstrate value. People often apply old metrics to new capabilities. The hosts gave the example of customer service handle time. When paying humans by the hour, minimizing handle time makes sense. But AI agents can engage customers indefinitely, potentially generating incremental revenue through extended conversations that would never have happened before.
MIT found that external partnerships reach deployment about twice as often as internally built efforts, with 67% success versus internal teams. External partners provide best practices guidance that internal teams lack. The most successful projects targeted back office operations like finance and supply chain rather than flashy front office personalization.
IT resists most
A Master B2B poll found 60% cite IT as the most resistant group to AI adoption, compared to 20% for sales, 20% for customer service, and 0% for marketing. The hosts suggested IT resistance may stem from being responsible for backend integration where problems surface. They may also have legitimate security concerns about AI accessing proprietary information.
However, the hosts cautioned against excessive risk aversion, comparing it to people in 1910 who would have banned cars for being dangerous. Speed limits and safety belts address risks while preserving benefits. AI requires similar balanced governance rather than blanket rejection.