In 2012, we cared that we used software. Today we care how we use it.
The difference is trajectory.
In the last decade, adopting software was the priority. Moving from on-premise to the cloud or digitizing a manual workflow promised productivity gains. Adoption was the finish line.
Today software is ubiquitous. Every salesperson uses a CRM & every engineer uses an IDE. The edge no longer comes from having the tool but from the specific path & manner in which that tool is used to achieve an outcome : a trajectory through software.
A salesperson creates a lead, enriches the lead, adds in information about the prospect in a particular way. That’s one kind of trajectory. A Q&A session with AI is another trajectory : how do I conduct a research project with AI on post-quantum encryption? What are the leading algorithms? Which companies are implementing them? What’s the timeline for quantum computers to break current encryption? Who are the experts I should talk to?
Tracking a user working through the day like a pinball ricocheting around a machine is tremendously strategic.
First, automation requires trajectories. To automate work, you must first understand the path of that work. In the past we hired consultants to map processes manually. Now AI agents can watch & record & understand these trajectories in real-time. AI learns by observing.
Second, optimization requires repetition. Trajectories provide the dataset for improvement. By analyzing thousands of passes through a workflow, AI identifies success patterns & failures & inefficiencies.
Third, trajectories become the new moat. The higher the resolution of the data, the more differentiated the AI product becomes, which increases vendor lock-in.
Fourth, company leadership benefits from understanding employee trajectories. We think we work together in one way, typically with some aspirational ideas. It’s another to truly understand the workflows in the field.
Fifth, trajectories are the basis for optimizing AI models through reinforcement learning or fine-tuning. Smaller specialized models trained on high-value paths replace massive generalists. Lower inference costs & higher accuracy lead to increased margins.
The strategic nature of trajectories raises the question of whether enterprises will negotiate the rights to their trajectory data when buying AI software, both to capture critical data & prevent lock-in. How those power dynamics play out will determine the pricing power for software broadly.
The companies that master these trajectories will define the future of work.