News & Content

The Great Rebundling

John Assalian, CEO · November 9, 2026

Walk into any two car dealerships and you will find two different ways of doing the same thing. Different forms, different scheduling habits, different rules about who calls whom. Decades of consolidation have not changed this. The variance survives because workflows are not designed. They accrete. Each one is the residue of a particular person solving a particular problem on a particular Tuesday, and then everyone after them doing it that way because that is how it is done.

Every technology wave made this worse, not better. ERP systems were adopted bespoke. SaaS tools were adopted bespoke, by individual teams, often without telling anyone. No-code tools let anyone build a workflow, so everyone did. The result is that a mid-sized company now runs on hundreds of small, undocumented processes, bundled together in ways that nobody fully understands, including the people running them.

AI is the first technology that can plausibly unwind this. The promise is a universal operating layer: one system that understands the work, holds the rules, and lets people act through it. Put the mess back together in a form that makes sense. The opportunity to build those operating layers, company by company and function by function, is one of the largest in software.

There is a catch, and it is the interesting part.

AI cannot do it by itself. The workflows are not written down anywhere a model can read. They live in habits, in Slack threads, in the way one person always double-checks the thing another person always forgets. A model looking at a company from the outside sees the tools, not the work. The bundling humans did was so particular and so complex that no amount of intelligence can infer it. Someone has to go in, find the workflows, and recreate the software around them.

History has a precedent. When factories first got electric motors, productivity barely moved for decades. The gains only arrived when engineers walked the floor and rearranged the machines around the new power source instead of bolting it onto the old layout. The motor was the easy part. The rearranging was the work, and it was done building by building, by people.

That is where we are with AI and work. The models are the motor. The rearranging is a job, and it is a large one. In that sense this is a true boom, not in the speculative sense but in the plain one. There is an enormous amount of work to be done, and real value at the end of each piece of it.

Retent does this work for marketing. We start there for a practical reason. Marketing has more tools, more handoffs, and more variance than almost any other function, and it already runs on the AI platforms where the operating layer belongs. We go inside, find the workflows the team actually runs, and rebuild them as one system the team operates, end to end. Then we move to the next workflow.

The bundling took decades. The rebundling will take years, and it will be done one organization at a time. We think that is the most interesting place to be working right now.

A new era of work begins now.