Vision
Vision
The Work That Isn't Strategy
Organizations adopting agents were promised that the human work left over would be more strategic. What's showing up instead looks nothing like strategy. It's specifying objectives precisely enough that an autonomous system won't wreck anything, holding the conditions steady enough for it to keep running, and deciding which outputs are worth keeping. That work is demanding and largely invisible, because everyone was told those hours would go to higher-order thinking. The people doing it often believe they're falling short, when what they're actually doing is developing a skill nobody has named yet.
The Work That Isn't Strategy
Organizations adopting agents were promised that the human work left over would be more strategic. What's showing up instead looks nothing like strategy. It's specifying objectives precisely enough that an autonomous system won't wreck anything, holding the conditions steady enough for it to keep running, and deciding which outputs are worth keeping. That work is demanding and largely invisible, because everyone was told those hours would go to higher-order thinking. The people doing it often believe they're falling short, when what they're actually doing is developing a skill nobody has named yet.
Two Domains, Two Phases
What Scientists Actually Do When the Lab Runs Itself
A self-driving lab ran 3,648 synthesis experiments with no one touching the bench. The months of human work beforehand — choosing which chemicals to include, turning "better nanocrystal" into three measurable properties, deciding how many early cycles to spend exploring rather than optimizing — is where the chemistry actually happened. When machines handle execution, the difficulty moves to specifying what they should try.
Who Owns What the Machine Wrote
At two German firms, professionals wrote their own rules for AI output: no-delegate zones for legal text, a deliberate pause before accepting high-risk material, peer review borrowed from code workflows. At a software company where AI contributions reached 90% of pull requests, human review coverage fell from 89% to 68%. When machines generate the work, the live question is what you are willing to put your name on.
The Working Taxonomy
Organizations treat AI deployment as a single job. Look at what's actually happening inside any deployment and you find three distinct activities, each with its own failure mode.
Specification work happens before the agent runs: making a task explicit enough to delegate. Boundaries, acceptable outcomes, recovery paths. In autonomous chemistry labs, researchers define materials, objectives, and experimental budgets before the automated cycle starts. The system operates inside that envelope.
Execution work is the agent performing the task. Merged code, resolved tickets, processed cases. This phase absorbs most organizational attention because it produces countable outputs.
Adoption work makes those outputs usable: reviewing results with enough context to accept or refuse them, detecting failures, maintaining the organizational capacity to recover when something breaks.
Execution scales fast. Specification and adoption require human judgment that doesn't compress the same way, and that asymmetry is where most deployment-to-value gaps originate.
Outside Perspectives
