Start from zero: how Chatham reengineers work for the age of AI

Most firms begin their AI work in a sensible place: give people access to capable AI models, prepare a secure data environment, and start building and automating. The mistake is treating those steps as a straight line to getting better results. 

As our CEO Matt Henry puts it: “If you apply AI to a broken process, you will run a broken process faster. You won’t necessarily make it better.” Investment goes up, and more work becomes automated. But the assumptions, handoffs, and unnecessary steps built into the original process remain. 

That insight led Chatham to Process Zero, a method for redesigning work from the desired outcome backward—before deciding what technology to build to support it. 

Build fluency—without mistaking it for transformation 

It’s only appropriate to start from the beginning. Since ChatGPT’s launch, Chatham has deliberately focused on applying AI to drive better capital markets outcomes for clients. Leadership began by educating employees about AI and giving them secure access to capable tools. Rather than prescribing use cases, the firm encouraged employees to learn where AI could improve specific outcomes and ask whether the existing workflow should exist at all. 

Chatham Vibes, the firm’s secure, collaborative application-building platform, accelerated that learning. The point was not to treat every prototype as a product. It was to expose what the technology could do, and identify where it might improve real work. 

With Process Zero fundamentals as digital scaffolding, Chatham Vibes builds fluency and expands what teams can create. Quite successfully at that, with greater aptitude over time—in the last four months alone, Chatham employees have built nearly 1,000 applications. 

Put Process Zero before the technology 

The conventional sequence in software development is to build literacy, improve access to data, and then build technology around the existing workflow. Each step can create improvement. Together, however, they can lock sometimes decades-old processes into a new system. The organization ends up doing the same work faster rather than asking whether it should be doing the work that way at all. 

Process Zero interrupts that sequence. It is a platform- and solution-agnostic method. Before a team chooses an application, model, or platform, it names the output the process must produce. It then strips the process to its essential inputs and redesigns the path between them. Only after the new workflow is clear does the technology decision begin. 

The backward step to zero is the discipline. It forces the team to separate the outcome from the inherited process used to create it. Steps survive because they are necessary, not because they are familiar. 

Front-line talent plays an important role because it understands the day-to-day work and the constraints surrounding it. A fresh-eyes team challenges inherited assumptions and handoffs. Subject-matter experts then test the redesigned process and certify that its output meets the required standard. 

Choose technology for the redesigned process 

Once the workflow has been redesigned, the technology and data foundation has a specific job: support the new process. The redesign reveals what information must be captured or retrieved, where AI adds value, and what systems should connect the essential inputs to the desired output. 

The answer is not predetermined. The right solution might use Chatham Vibes, an existing internal system, vendor software, a large language model, conventional automation, or no AI at all. Process Zero determines what the process needs; it does not prescribe the platform. 

Whatever technology is selected, a prototype is not a production system. It must be tested, secured, permissioned, integrated, monitored, governed, compliant, and auditable, with a named owner and subject-matter expert responsible for the final go or no-go decision. When there’s a go, based on explicit acceptance criteria, the process and supporting tech also needs a named owner accountable for performance after launch. 

Operationalize, measure, and return to zero 

Chatham monitors three things to ensure success long term: whether the output continues to meet the required quality standard; whether unnecessary inputs or steps are creeping back into the process; and whether the output itself is still the right one. A failure in any of those areas sends the process back to the appropriate stage of Process Zero. 

That return to zero matters. Without it, even a well-designed process can accumulate complexity and become the next generation’s inherited workflow. 

The advantage is the process 

Most firms have access to the same leading models and, increasingly, similar application-building tools. Like access to IT in the ‘90s, neither is likely to be the moat today.  

Bolting automation onto inherited work is the available shortcut, and a trap. Rebuilding the process with AI-native support is the opportunity, and the two start from different premises. Starting from zero does not mean starting over. It means refusing to let today’s workflow—or a preferred platform—dictate tomorrow’s system.