AI Adoption Fails 95% Of The Time. Small Business Leadership Is Why - Forbes

Companies have invested $30-40 billion in Generative AI during the last two years yet only 5% of AI pilot projects produce an identifiable ROI. According to MIT's Project NANDA, 95% of all money invested in AI by companies is invisible from a profit and loss perspective. For the small business owner investing $200/month in an app or service expecting it will self-fund, this is likely worse than for the CFO at a Fortune 500 Company. Developing a viable small business AI plan begins with understanding your business first. How you operated prior to AI arriving will determine most of why your small business may fail to implement AI.

Leadership Hasn't Caught Up To The Tools

New research, published on July 22nd, 2026, by ManpowerGroup Talent Solutions came to the same conclusion as the other studies. The study was developed by Everest Group and included the responses from 80 C-suite and senior talent leaders. Of those surveyed, only 3% reported that their current leaders are fully prepared to lead an AI-enabled team. In addition, McKinsey’s 2025 State of AI Survey found nearly the same result; only 1% indicated that their organization has achieved full maturity in its AI Strategy. While 88% of organizations have now implemented some level of AI within their businesses, the readiness of leaders to leverage AI is still low.

Boston Consulting Group reported that 70% of AI transformation efforts in business organizations do not meet leaders' expectations for results. They identified organizational culture as the root cause of this gap rather than the technology itself. Organizations that put at least 10% of their AI budget into training and change management were one and a half times more likely to succeed than organizations that didn't, according to BCG. AI does not develop new processes, it enhances existing ones by running them faster than people can write or enter data.

What The Successful 5% Do Differently

The majority of the 5% who successfully used AI Process Automation identified patterns in their implementations. The process automation pilots implemented by those companies were limited to repetitive daily processes. These included things such as processing invoices, scheduling and completing back office functions which already have an established repeatable process. Although a broad based ChatBot may appear flexible on a sales page, they will be difficult for the average business to implement because it does not have the ability to understand or adapt to the specific workflow of that company. Using the same type of AI-based software system as a Fortune 500 company is no substitute for creating the workflow necessary to use this technology effectively. Therefore, AI implementation for small businesses can only provide significant returns on investment if the workflow has been previously documented.

The Founder Dependency Problem

Most small-business owners will pass over this step. According to a study conducted by the National Association of Insurance Commissioners (NAIC) regarding the role of founder dependency in small business. 71% of small businesses depend on either one or two individuals for their overall success. The NAIC study was referenced in nationwide’s “Key Person Risk” research. When lenders consider the potential risks associated with acquiring a small business they assess whether the business has sufficient controls, processes and redundancy for day-to-day operations if the founder were to leave.

The Small Business Administration (SBA) lenders perform their own dependency assessment when reviewing an acquisition transaction. Even if a firm passes the SBA lenders’ dependency assessment, there are no guarantees that the firm will receive full funding. Instead, the lender may provide less than what would be expected under normal circumstances. Additionally, the lender may require a longer transition period with the founder remaining on payroll during which time the lender may also include life insurance coverage to ensure continued operation of the business if the individual responsible for running the business leaves.

One Founder Who Closed That Gap

Anthony Godley is acutely aware of that chasm. He founded Logix BPO alone; he had just one client at inception. He was able to grow Logix BPO into an outsourcing company employing over 1,000 people when he became its chairman in 2025. In a written statement, Godley noted: "The biggest barrier to AI adoption isn't technology. It's founder dependency. If every important decision still comes back to the founder, AI will only expose that bottleneck faster."

Godley said, "They already have disciplined processes and leaders who don't have to make every decision themselves. AI amplifies operational maturity. It doesn't replace it." Make decisions first, then use software. Document each and every approval and escalation that will need to come from you. After that document how each of those approvals is made. Document all of the "judgment" calls a spreadsheet or AI will never be able to make. The documentation of this workflow will provide an AI with enough context so that it may assist you with taking some work away from your own workload, as opposed to just producing additional output for you to verify. For business owners looking to determine what their ROI would be on implementing an AI system in their small business, the key is to start by documenting a real process that has quantifiable results.

Start With The Decisions, Not The Software

Ask yourself three questions before you consider purchasing the next AI product. Who else in your organization has the authority to make this decision about how to use AI? Is there a definition of success when it comes to using AI on this task that is documented somewhere so someone other than you could read it? Will the process continue to function even if you are out of town for a week? If your firm is unable to answer these questions today, then it will be better able to answer them with an AI subscription. The subscription will simply provide more tools for longer to avoid asking these questions.

Identify someone at your company who has the least documentation for their most critical tasks and cross train them on those tasks by the end of the month. Not because and AI vendor told you so, but rather because firms consistently achieving better than a 95% failure rate have known “who else can do this” prior to buying the first AI seat. Once you have identified that person and answered in this manner, next identify a method to measure or determine your small business’ AI Return On Investment (ROI) as well as the cost of the tool purchased and the purchase will be the easiest part of the project versus being the entire project.

As long as the 95% failure number keeps getting quoted as an AI problem, it will keep being treated the wrong way. It's a leadership problem wearing an AI price tag. Those companies in the remaining 5% didn't always possess a superior algorithm. They already were building a business based on a system of writing, not just one individual's personal experience, and that process-oriented culture enabled their ability to implement successful AI. Implementing AI was the simple part they decided to add at the end.