Impact recognition
We are honored to be named an Inc. Power Partner for the third consecutive year. This recognition is achieved by B2B companies that help entrepreneurial businesses grow, and it rests on feedback from the clients those companies serve.
That distinction matters most to us. Anything we write about our partnerships is marketing. What our clients say is evidence. With an average client tenure of more than three years, and our longest relationship lasting over twelve, there has been plenty of time to form their opinion.
The award builds on five consecutive years on the Inc. 5000 list of America's fastest-growing private companies. Our relationship with Inc. now extends past the lists. COO Don Gregori is a member of the Inc. Leadership Forum and a contributing author at Inc.com, where he writes about leading teams and the realities of building software with AI. His recent piece, "One Feature, 11 Human Interventions," examines what it takes to build with a software factory: an assembly line of agents building a product from a single prompt. You can read his work at inc.com/author/don-gregori.
A year of doing more with less
Most every company we spoke with this year wanted the same three things: lower costs, more output from the teams they already have, and a credible plan for AI. The order changed by industry. The pressure did not.
Senior client stakeholders asked to ship faster while keeping or reducing headcount. And every client was on an AI path, even if they were stalled at the AI starting line.
Moving fast with AI and no guardrails creates risk. Governing too cautiously impacts output, which is a different kind of cost.
Why most AI projects never reach production
By the estimates cited in a 2024 RAND Corporation study, more than 80% of AI projects fail. That is roughly twice the failure rate of IT projects that do not involve AI.
The model is rarely the problem. RAND interviewed experienced data scientists and engineers and traced most failures to a few issues: the data needed to train an effective model did not exist, the team chased the newest technology instead of the user's actual problem, or the infrastructure to deploy and govern models was missing.
Every one of those issues is visible before the coding starts.
Here at First Factory, we haven't been racing to get AI products out for bragging rights; we've been speeding to gain pivotal expertise in the domain so that we can launch AI projects with significantly higher percentages of success.
Certifying the whole company, not just the engineers
More than 20 people at First Factory now hold Claude certifications. By the end of next year, we expect every person in the company to be certified, including recruiters, designers, HR, and operations staff.
The reasoning is practical. A company that asks clients to adopt AI responsibly should hold itself to the same standard. When the people scoping a project, staffing it, and supporting it all understand what the tools can and cannot do, there's less friction and greater alignment. This expedites results.
This October, First Factory became a Claude Select Partner. The designation reflects that investment and gives our clients a partner with direct expertise in the platform many of them are building on.
The pressure on technology budgets will not ease next year, and neither will the expectation that AI pays for itself. We remain focused on helping companies be successful in spite of those forces.
Sources
- RAND Corporation, The Root Causes of Failure for Artificial Intelligence Projects and How They Can Succeed, August 2024.
.avif)
