Home / AI change management
AI transformation
AI change management is the work of moving people, roles, and rules to a new way of working once AI agents take on part of the job. Cameron LiButti led this change at Bidview Marketing, the agency he founded. He moved the team to an agent-first operating model, where humans own the client relationship and AI agents own the execution.
The model
AI change management at Bidview started with the operating model. Every job now runs on the same split.
The system behind it is a multi-tenant AI platform the team calls the harness. One company-level control plane holds the method. Seventy isolated client environments run on top of it.
The change
Bidview's methodology, prompts, and interpretation rules live in one company-level control plane as versioned IP. When the process changes, the version changes, and every client environment runs on the new rules. AI tools and employees work from the same written method.
AI agents do 80% of the execution. An employee reviews the work before it reaches a client. That review is the last 10% of every job, and it is where trust in the AI output is earned.
Cameron did not draw a new org chart where humans and agents had to learn to work side by side. He changed the engine of the company: the system that pushes tasks and reviews data. Admin work moved into that engine, employees moved to reviewing its output, and Cameron updated the job descriptions to match.
An AI-enabled operating playbook trains account managers on the new workflows. New employees learn the agent-first model from day one, so adoption does not depend on who happened to be in the room.
The impact
The benefits showed up in capacity and speed. Every analysis cycle moved one step faster, and the team now takes in more competitor data than before.
The timeline
The hardest part of the implementation was the technology itself. Cameron started the first version of the system in February 2026. It was not capable of evolving on its own, and a system that cannot improve itself would not keep up across 70 client environments. So he tore it down.
The second version started in March 2026 and was built to learn. The method lives in one versioned control plane, and the system takes in employee feedback every week. By June 2026 it was working across the company.
The people
Employees did the execution by hand. The work depended on each person remembering the process.
AI agents do the execution. Employees review the output, keep the client relationship, and make the decisions that stay human.
An AI-enabled operating playbook trains account managers on the new workflows and the review gates.
Positively. The AI technology was built to assist the team without forcing a new way of working, so adoption did not depend on people changing their habits.
The team talks about the system every week. It is built to learn from employee feedback, so the people closest to the work keep improving it.
Clients care about outcomes. Cameron explains the system to clients who are curious. Most do not ask.
Governance
Every prompt and rule is under version control, so a change to the process is tracked like a change to code.
No agent output reaches a client until an employee reviews it.
Each client runs in its own isolated environment, so one client's data never mixes with another's.
Lessons
These are the lessons from AI adoption at Bidview that Cameron shares with leaders and organizations planning their own change.
At Bidview, the AI technology changed the engine that pushes tasks and reviews data. Teams kept their roles and their client relationships. The team reacted positively, and adoption did not depend on persuading people to change their habits.
Bidview's first version could not evolve on its own, and Cameron tore it down after about a month. Leaders planning AI adoption should set aside time and resources for a rebuild, because the first version teaches the organization what the second one needs.
The team talks about the system every week, and it is built to learn from that feedback. The people closest to the work keep improving it, which turns employee experience into a lasting capability for the organization.
Bidview measures the impact of AI in two ways: how many clients one account manager can handle, and how often analysis runs. Both are productivity numbers that leaders and stakeholders can check every month.
Every piece of agent output passes a human review before a client sees it. That review gate is what lets an organization trust AI insights at scale and keeps a culture of accountability in place.
The operator
Cameron did this as the operator. He made the architecture decisions and rewrote the job descriptions himself, so the AI tools and the roles they changed had one owner.
Before marketing, he spent about five years as an MEP consulting engineer in Chicago. He still starts every change the way he started a boiler retrofit: find out what is broken and what it costs before anyone buys anything.
Questions & answers
AI change management is the work of moving people, roles, and rules to a new way of working once AI agents take on part of the job. It covers who reviews agent output, how the method is written down, and how the team is trained.
He moved Bidview Marketing to an agent-first operating model. Humans own the client relationship. AI agents own the execution. Work runs on a 10/80/10 split: 10% human strategy up front, 80% agent-controlled execution, and 10% human review at the end.
Account managers now handle double the clients. Every analysis cycle moved one step faster: quarterly work became monthly, monthly became weekly, and weekly became daily. The team also takes in more competitor data than before.
At Bidview, work started in February 2026. The first version had to be torn down because it could not evolve on its own. The second version started in March 2026 and was working by June 2026.
Positively. The AI was built to assist the team without forcing a new way of working. Cameron changed the engine that pushes tasks and reviews data, and left the org chart alone.
At Bidview, admin work moved into the AI engine and employees moved to reviewing agent output. Cameron updated the job descriptions to match, and an AI-enabled operating playbook trains account managers on the new workflows.
Clients care about outcomes. Cameron explains the system to clients who are curious, and most do not ask. The team talks about the system every week, and it is built to learn from their feedback.
Every prompt and rule is under version control. Human review gates sit between agent output and the client. Each client's data stays in its own isolated environment.
At Bidview, the founder led it. Cameron LiButti made the architecture decisions and rewrote the job descriptions himself, so one person owned both the AI tools and the employee roles they changed.
Cameron starts with an audit: what is broken, what it costs, and which fix pays back first. At Bidview, the method then lives in one control plane, so AI agents and employees work from the same versioned rules.
At Bidview, an AI-enabled operating playbook trains account managers. Employees learn to review agent output before it reaches a client, because review is their main job in the agent-first model.
Owners and leadership teams who want to run their business agent-first. Contact Cameron by email at cameron@bidviewmarketing.com or through the contact page.
Want to learn more about agent-first businesses? Contact Cameron to talk through what the move looks like for your company.