AI cannot run on chaos.
It can’t run on handshake deals, tribal knowledge, and six layers of management that exist to translate between each other. There is nothing there to automate. Seamless builds the structure first, so the AI has something to run on.
Billion dollar decisions are moving through Slack channels.
You didn’t have this problem at five projects. You have it at fifty.
- Executive asks for portfolio status. No one has a real answer.
- Your PMs spend 40% of their time chasing updates, not running projects.
- A coordination failure becomes a crisis by the time it reaches leadership.
- By the time issues surface, the mitigation options are limited.
- 90% of projects are delivered late to the customers waiting on them.
- Every new site invents its own version of the process, because the process can’t travel to a new site without a person carrying it there.
- The structure that would fix it is sitting in your best people’s heads.
Adding AI to this doesn’t fix it. It amplifies it.
The solution is simple, not sexy.
Your tech and construction teams don’t have a shared standard to work from, and neither does the AI. So build one.
Find your top performers
Find the people who deliver results. Ideally not the managers, but the ones performing the work.
Document what they do
Have your top performers provide the runback for what they do. What are all the steps, one through ten? What are the templates for each step, which stakeholders are involved at each step, and what are the risks to be aware of?
Then find the gaps and define them The step everyone skips
Find the pain points, the risks, and the unanswered questions that always exist amongst large, multidisciplinary teams. Ensure that the structure you create is holistic, and at least creates an opportunity for a conversation to happen.
Build the backend system for your operations
Build a roadmap for your teams that doubles as a reporting layer. Teams review and update it throughout the lifecycle, so it becomes part of the work instead of a task on top of it, and accurate status reporting happens on its own. Teams are not going to keep referencing a runbook or a playbook, so the process information is built into the reporting itself. We build in Asana by default. Your teams already know it, and that is the difference between a system people open and one they ignore.
Then layer AI on top
Once the workflow is standardized within a system, automation can be layered on top effectively, improving employee productivity, enhancing work-life balance, and creating opportunities for OpEx savings.
Build the foundation. Then let AI run it.
- Your best person taking leave stops being a schedule risk.What they know lives in the system, not just in their head.
- Administrative load decreases.Project managers focus on reducing risk instead of spending their time in meetings and building PowerPoints.
- You are less reliant on expertise, since it lives within the system itself.You can expand hiring to any competent hire, not just experts.
- Reduced risk across the portfolio.The system closes the gaps and keeps every stakeholder looped in.
- You are able to remove layers of inefficiency.Work moves without needing excessive oversight.
- You onboard in weeks instead of quarters.New people work inside a defined process on day one.
Start with the Systems Scan.
Audit
Systems Scan · 3 weeks · credited in full toward Offer 1
How your operations actually function today. Where the friction lives, where the risk sits, and what has to be true before AI can work here. You get the diagnostic, the top three breakdowns, and a go or no-go.
Roadmap
Operating System Roadmap · 6 weeks
Full audit of the program, the risks, and the dependencies. The roadmap defines what the system should handle and what your teams need to own. Routing, status and visibility are built into the system, so your people are able to focus on judgment calls, relationships and delivery instead of managing information.
Design
Operating System Design · 3 months
The complete system design, ready to deploy. The process lives inside the tool your teams already use, so as they update their work throughout the lifecycle they are able to see what comes next. Includes the role structure, the tool stack, and the governance that defines how work routes and who approves what.
Execute
2 to 3 months · declining monthly structure
The blueprint becomes operating reality. Tooling, integration, leadership training, and office hours through rollout. The declining price is intentional. The goal is your independence.
Automate
The Automation Layer · final phase, scoped at steady state
The reason for every phase before it. Once the system is live and stable, we automate the operation itself. The repeatable work stops depending on people.
Built at Meta. 80 data center builds, and the reporting stopped eating the week.
Meta was running 80 active data center construction sites across the US. Above each site’s program manager sat a layer of TPMs, each covering roughly 20 projects, whose job was making sure teams hit milestones and followed process.
The process documentation existed. Nobody read it. So the TPMs became the process: constant check-ins, manual status tracking, chasing updates across 80 moving parts. Good people spending their days moving information between other people.
Hannah’s team put the operational guides directly into Asana. Every milestone a team hit, the steps were right there. External teams updated their own boards. Those updates rolled into a live portfolio view leadership could open any time, without asking anyone.
Status chasing stopped. Reporting became automatic. The coordination layer got to spend its time on judgment instead of information. Piloted across the full portfolio, field to executive.
Credibility is the whole point.
Hannah Garrett spent the last 10 years building commercial and mission critical projects across six companies, including Meta. The kind of programs where one coordination failure costs millions and decisions can’t wait until Monday.
We’re not an AI company that picked up operations. We’re an operations practice. We know where AI helps and where it makes things worse, because we’ve watched both happen on real projects.
Ford ran this experiment in reverse. They leaned on AI quality inspection, then brought back roughly 350 engineers in June because the systems weren’t catching defects. Their VP said it plainly: AI is only as good as the information you train it on. They got to the right answer. It just cost them the defects and the year to get there.
Start before you’re forced to.
Do not press the panic button. Future success takes strategic, methodical planning. Take the chaos out of it and create the structure AI needs to succeed. If you are not laying the foundation now, you will not be ready to implement the models that are evolving as we speak.
Book a Systems Scan call