Building autonomous AI agents. Notes from the team building Gravity. AI workflows, the future of recurring work, and what we learn along the way.
Production AI agents need updates. Models improve. Prompts get tighter. Tools get added. The team finds a way to make the stopping rule clearer. The question is not "should we update" but "how do we update without…
Read post →
The single hardest non-model problem in agent engineering is state. The model is stateless. Every other part of the system that gives it the illusion of continuity, of memory, of resumption, is your code. Get it…
Read post →
Code without version control is a hobby. Prompts without version control are a liability. The reason most agent prompts produce silent regressions in week six is not that the prompt got worse; it is that nobody can…
Read post →
Prompt engineering for AI agents is not the same craft as prompt engineering for chatbots. A chatbot prompt shapes one response. An agent prompt shapes a loop: the model picks a tool, reads the result, decides…
Read post →
The intuition that "more agents will do better than one agent" is wrong more often than it is right. Most production multi-agent systems exist because the work has genuine boundaries (different access controls,…
Read post →
Most AI agent failures in production are not model failures. They are integration failures. A webhook arrives twice and the agent acts twice. An OAuth refresh fails silently and the agent runs unauthenticated. A…
Read post →
The model is not the safety system. The model is the part you are deploying. Every control that protects users, data, and the bill from the model goes around it, not inside it. This guide is the runtime controls…
Read post →
Time-to-first-action is the most honest deployment metric for AI agent platforms. Marketing pages quote "deploy in minutes" without saying what counts as deployed. This benchmark sets a fixed task, an inbox-triage…
Read post →
Data residency is one of the silent gating items for enterprise sales of AI agents. The product can be perfect, but if the prompts leave the EU, the deal dies. This guide is the architecture playbook for…
Read post →
An AI agent's audit trail is the difference between "the agent took an action" and "we know why the agent took that action." It is what enables incident response, compliance audits, and the kind of post-hoc analysis…
Read post →
The honest pitch for AI agents to a SaaS founder is not "scale your team." It is "stop dropping the recurring five-to-fifteen-minute tasks that compound into churn." Most SaaS founders I talk to do not have a…
Read post →
Recruiting is the AI agent use case where the operator advice and the regulatory advice point in slightly different directions. The operator wants the agent to shortlist, screen, and schedule. The regulator wants…
Read post →
Real estate is one of the few industries where a single AI agent can be the difference between converting a lead and losing it to a faster competitor. Inbound buyer leads are perishable in a way that almost no other…
Read post →
For most small-to-mid marketing agencies, AI agents are the difference between a 12% net margin and a 28% net margin on the same retainer book. Not because the agency hires fewer people, but because the agents soak…
Read post →
Lawyers face the steepest constraints of any profession deploying AI agents in 2026 and the largest potential upside. The constraints are real: client confidentiality under Model Rule 1.6, competence under Rule 1.1,…
Read post →
The honest job description for a freelance designer is "two roles in one person." Half the week is design work; half the week is studio operations. Intake calls. Scope clarification. Proposal drafts. Asset packaging.…
Read post →
Open with margin, not vibes. A DTC store doing $1M ARR at a 65% gross margin and 7% net has roughly $70k in net profit and zero room to add headcount. An AI agent that recovers 2% of abandoned carts on a baseline…
Read post →
Content creators in 2026 face a paradox. The tools to make and distribute content have never been more powerful, and the audience tolerance for AI-generated work passing as human has never been lower. The creators…
Read post →
Consulting is a profession built on judgment, but a typical engagement is filled with the supporting work that produces the conditions for judgment: research, synthesis, slide drafting, meeting prep, status…
Read post →
Accountants and bookkeepers face two pressures pointing in the same direction. Clients increasingly expect monthly close in five business days instead of fifteen. Margins on transactional bookkeeping are compressing…
Read post →
Most security guides written for large language models stop at the prompt boundary. They assume a single completion, no tools, no state, no autonomy. That model has not described production deployments for at least…
Read post →
Most engineers transition from testing deterministic services to testing agents and discover their old reflexes do not work. Unit tests assume identical inputs produce identical outputs. Agents do not. Integration…
Read post →
The first time I shipped an agent without proper observability I did not notice quality degradation for nine days. Token costs were stable, latency was fine, error rates were nominal. The agent was answering…
Read post →
For most of 2024 and the first half of 2025, AI governance for agents was a tomorrow problem. By mid-2025 it had become a this-quarter problem. The EU AI Act began entering force in stages, the NIST Generative AI…
Read post →
An agent that handles errors well looks identical to one that does not, right up until the day a tool returns a 500 in the middle of a refund chain and the customer ends up with the money back but the access revoked.…
Read post →
This is not a primer on AI agent pricing models. The taxonomy of per-token, per-task, per-agent, and capability-based pricing already lives at AI agent cost models explained. This piece is the operational sibling.…
Read post →
Workato and Gravity get put on the same shortlist when a buyer types "AI workflow automation" into Google and the search engine hands back both. The shortlist is misleading. Workato is an enterprise iPaaS. Gravity is…
Read post →
Tray.ai and Gravity both ride the AI automation wave, but the buyer and the building experience are different. This comparison helps decide which category fits the work you have.
Read post →
Relay and Gravity both target the "AI drafts, human approves" pattern. The category framing is different. Relay puts the workflow first and the AI inside a step. Gravity puts the outcome first and approval inside the…
Read post →
Pipedream and Gravity get lined up a lot because both can "take a Slack message and do something useful with it." The interface is different, the buyer is different, and the unit of work is different. This is a…
Read post →