AI Gains Traction When Companies Reduce Friction, Turn Managers Into Workflow Owners

For many companies, rolling out AI looks like a standard IT update: buy the licenses, run a quick prompt-training session, and check the box. But a few months later, usage flattens, and projects stall. Looking for a solution, executives dump the initiative onto the most tech-savvy person in the room without giving them the time or structure to actually manage it. That person dives in, realizes the scope is essentially a second full-time role—learning the tools, training colleagues, building automations, managing upkeep—and produces no results.

Patrick Schaber, an AI transformation leader and Fractional CMO, spends most of his time in that messy space between initial adoption and true innovation. As Founder and CEO of Patrick Schaber AI and Marketing Consulting, he helps leadership teams turn high-level AI ideas into practical workflows that employees can use every day. Previously, he led enterprise-wide AI adoption as Head of Marketing and AI Implementation at GovDocs, and has a record of building inbound engines that generate the majority of the pipeline with zero ad spend.

His view is straightforward: the real work of AI is less about saying "we implemented AI" and more about a true understanding of what AI gives to an organization. "The companies that succeed with AI are the ones where leaders have actually done the work themselves," Schaber says, "because they understand how to identify the right workflows, the right architecture, and where the real value is.

• People over processors: The technology is putting more emphasis on human capital than on tooling. "The more AI evolves, the more investment in people is going to matter. There are some places where it can replace a workforce, but for the most part, most companies are not going into it to replace people. They're going into it to make their people better at what they do."

• Breaking the bottleneck: Without structured training, Schaber sees that adoption fractures. "Take a group of 10 people. There are going to be two people that really go crazy with it, and then everyone else is going to fall behind. Then there's going to be a bottleneck where the people that are using it are dramatically creating more output, and the other people aren't keeping up."

So, who actually has to smooth out that friction? In practice, the job frequently falls to middle management. Directors and VPs are now expected to identify and automate a large share of their group’s tactical work while also helping those same employees move up the value chain. Organizations can preserve institutional knowledge by pairing strong but non-technical leaders with dedicated internal or external support, instead of expecting every manager to become a prompt engineer overnight.

• Caught in the middle: For Schaber, juggling both automation and team support fundamentally changes what it means to run a team. "In order to realize the benefits of AI, those middle managers, those director-level, vice president-level, they're going to have to be able to identify where that 40% of automatable work is. They're going to have to understand how to make that work, and then they're going to have to understand how to get their people skilled up."

• Coaches over coders: The answer is not to replace non-technical leaders but to surround them with the right expertise. "At a large company, I think you're going to have an AI center of excellence or an AI team that's going to be able to be a partner to some of these middle managers. You could have a very skilled vice president of marketing who has done wonderful things for your company. They may have no technical skills, and that's okay."

For smaller companies without a formal AI team, Schaber often plays that support role himself, joining leaders a few hours a month to help them work through where the technology fits and what’s realistic to automate. The playbooks are still being written, and even the most eager teams are flying blind. As these new workflows spread across a business, many leaders find themselves reckoning with the legal and security implications of their tools. Without clear guidelines, well-meaning employees trying to save time can inadvertently feed company data into public models. A deliberate infrastructure allows companies to protect their IP and comply with internal governance requirements, while still giving teams room to learn.

• Sandboxes over silos: Unregulated use poses practical risks to proprietary data. Recalling a team that fed proposals and call transcripts into a public model, Schaber notes, "Most people are starting to realize at this point that using free public tools is educating a broader model, and you're putting all your information out there." Building those boundaries often means putting approved toolage in place or defining pathways that employees can use. "You want to encourage that exploratory behavior, but put the guardrails up so they understand the risks."

• Plumbing before prompting: Even well-designed sandboxes and content agents will struggle to deliver value without basic data hygiene. AI is an operational amplifier; it works best when the underlying processes are already clear. To get the most out of a new tool, Schaber advises treating data hygiene as a basic prerequisite. Before rolling out a new workflow, teams need to know where their data is stored and whether their processes are documented. He approaches this work as a pre-implementation audit. "There should be a checklist of things you have to go through before you even think about putting AI in. How good is my data? How easy is my data to work with? Is my process documented?

The value of these secure sandboxes often becomes clearer when looking at the reality of corporate AI literacy—which, in Schaber's experience, can be lower than tech insiders assume. Meanwhile, many executives are reading success stories about competitors, understanding and realizing the benefits, and asking why their own projects haven’t produced similar results. Traction is far more likely to sustain itself when leaders can easily calculate the ROI of a specific workflow rather than betting on a large, abstract program.

This is the heart of Schaber's AI adoption. As middle management wrestles with the challenge of automation and team leadership, it's up to leaders to understand exactly why they are adopting AI in the first place. What problems does it solve? What infrastructure exists to support it? How does it, in turn, support that infrastructure? If the right processes and people aren't in place, for Schaber, then that organization won't find success. "There are the security parameters, the data, a whole slew of things they have to get their act together on before they even think about putting AI in."