How AI Changes the Game
For years, custom software was a luxury. Most organizations bought the same applications as everyone else, then adapted their work to fit.
AI is changing that calculation. Software built around a particular organization is becoming practical for more businesses. The question is shifting from which tools to buy to how the business should work.
CYCLE helps you make that shift: understanding your organization, building the systems it needs, and guiding their evolution over time.
01 / The shift
Your way of working deserves more than a workaround.
One application holds your contacts. Another tracks projects. A spreadsheet connects the two. The person who knows why a certain client is handled differently is the only place that rule exists.
Shared software made powerful tools affordable. But serving thousands of companies meant designing for what they had in common. The details that made your organization different were often left outside the system.
The inherited model
Buy separate applications. Move information between them. Ask people to remember the exceptions.
The emerging possibility
Model the organization itself. Connect the tools that serve it. Give people and AI a shared foundation to work from.
What would you build if your current software were no longer the constraint?
AI can accelerate research, prototyping, implementation, and revision. That gives a smaller team room to solve more particular problems. It also makes the decision about what to build more consequential.
02 / The Company OS
A system that represents your company.
A Company OS is a shared operational foundation for your data, relationships, knowledge, rules, and workflows. It connects what your company knows with what your people need to do.
It can coordinate existing tools. Accounting, email, and payments may stay where they are. Your own operating model becomes the center.
Operations view
Permissions & approvals
Data Knowledge Business rules
- Accounting
- Payments
Operations view
The work, and what comes next.
A team's job queue, current status, and the information needed to move work forward.
An exception outside the team's authority goes to the appropriate person for approval.
Each view follows the person's role and permissions.
One foundation. Different ways to work.
Operations might need a job queue, leadership a cross-company view, and a client a simple project portal. Each can use the same underlying information with different permissions.
Interfaces and AI models can change. The more durable asset is the knowledge of your business: how information is organized, which rules apply, and who can act.
Inside the Company OS →03 / What becomes possible
Start with the work that doesn’t fit.
A Company OS develops around real needs. These are a few examples of the problems that can give it shape.
Knowledge that stays
Make the history behind a relationship available to the next person working on it. Bring documents, decisions, and context together with appropriate access controls.
A question it could answer
“What did we agree with this client, and why?”
Operations that connect
Carry a customer’s request through a quote, an order, inventory, delivery, and follow-up. Let each team work from the same understanding of its status.
A question it could answer
“What needs to happen before this order can move?”
AI with a defined role
Let an agent prepare a proposal from approved prices or retrieve permitted records. Keep consequential commitments subject to human approval.
A question it could answer
“What can be prepared now, and what needs a decision?”
04 / Direction & transition
Building the application is only part of the work.
The harder questions are specific to you. Which processes create value? Which are leftovers from old software? What knowledge has never been written down? Which decisions need a person?
A technically plausible answer can still be wrong for the organization. Direction means knowing enough about the business to make that distinction, and taking responsibility for the decisions that follow.
You cannot pause a business while replacing the systems that run it.
Customers, employees, data, and obligations already exist. We plan change around that reality: phased releases, reconciled records, time to learn, and a route back if a transition does not work as expected.
Understand
Learn how work actually happens, including its exceptions, relationships, and unwritten rules.
Model
Give that understanding a structure: information, workflows, permissions, and responsibilities.
Redesign
Preserve what makes the business distinctive. Reconsider what only compensates for old limitations.
Build
Develop and test the foundation, integrations, interfaces, and bounded AI capabilities.
Migrate
Move data, people, and workflows in phases while essential operations continue.
Direct
Keep improving as people learn, the business changes, and new possibilities emerge.
05 / CYCLE, since 2007
Experience that accumulates.
CYCLE began in February 2007 with design, websites, and visual communication. Our work grew into custom applications, databases, and systems that help organizations operate.
Today, that combination of design, technical knowledge, and organizational understanding informs how we approach AI. We work as a small senior partnership, directly involved in the decisions and the work.
The relationship matters. Over time, we learn why a process exists, which exceptions deserve care, and what a client values beyond speed. That context makes our direction more useful.
Meet the practice →Current & past clients
Our work spans professional services, education, nonprofits, finance, and founder-led businesses.
06 / A few useful questions
The questions behind the decision.
Do we have to replace everything?
No. Begin where your current tools constrain the work. Keep useful services and connect them to a coherent foundation. A proprietary system should earn its place by addressing a specific need.
If AI can code, why work with CYCLE?
Producing code is one part of a much larger decision. We help clarify the problem, design the system, evaluate what gets built, and plan how people will use it. Responsibility for that work stays with us.
How do we keep control of our system?
Ownership, access, documentation, and handoff arrangements belong in the engagement from the beginning. We plan for understandable systems, explicit permissions, and replaceable services so future choices remain practical.
Our continuity principles →Where would we start?
With a conversation about what is difficult today. We look for a bounded first problem, a senior person who can own decisions internally, and a reason to believe a tailored system would improve the work.
Thinking, by CYCLE
A changing technology needs a considered response.
Neurotransmitter is our publication on AI, software, and the changing organization. Its essays and working papers explore the assumptions behind our work—and give us language for the decisions ahead.
Read Neurotransmitter →Start with your organization
What are you trying to understand, preserve, or change?
We work best with small and mid-size organizations whose particular way of working matters, and with leaders ready to take part in shaping what comes next.
You don’t need a software specification to begin. Tell us where the current system stops making sense.