What it takes to put an AI agent to work
An AI agent can be a wonderful colleague for routine work. Here is what we have learned about the system it needs around it to be dependable, in any industry.
ReadWhat closed sales taught us about following up
Before building an AI teammate for insurance agencies, we studied an agency’s sales history in depth. What it showed about persistence and timing is encouraging for any sales team.
ReadHow an AI agent confirms appointments, and when it steps aside
A look inside Remi, the agent that confirms appointments for clinics, and the thoughtful details that make it comfortable for patients and useful for the front desk.
ReadDesigning outreach that people can trust
Email outreach works best when it is relevant, well-paced and respectful. Here is how we designed an AI system to help a sales team do exactly that.
ReadFrom a text message to a planned route
How an AI agent helps a waste-services company take bookings in the customer’s own words, and keep every driver’s day running smoothly.
ReadHow a legal AI agent earns an attorney’s trust
Legal work asks for accuracy, confidentiality and careful review. Here is how we designed an AI agent for law firms around those expectations.
ReadGiving AI agents a process the whole team can see
For a research consultancy we built two AI agents, a project manager and an epidemiologist. The most valuable idea was making their way of working visible and editable.
ReadSubagents, Agent Teams, and Durable Agents
Claude subagents and agent teams spin up a team for the session in front of them. Durable agents accumulate knowledge of your product over months. The lifecycle is the difference.
ReadHolographic vs Hindsight Memory
Our agent team runs two memory systems. Holographic for every engineer doing the building; Hindsight for the CTO and principal engineers doing the planning. Why the split — and what each is actually made of.
ReadProfiles
Pick a profile and the agents adapt — engineer peer review for builders, strategic partner for product owners.
ReadThe Agent Wiki
A living knowledge graph your agents read from before work and write to after shipping. Inspired by Karpathy's LLM wiki — built for progressive disclosure, not human browsing.
ReadAgent Performance Reviews
Every week, our agents are rated on ticket outcomes and the findings are written back into their memory — so each sprint starts from lessons learned.
ReadDaily Standups
The CTO agent writes a standup every morning — what shipped, what's in flight, what's stuck — so the team walks in with the board already read.
ReadYour dispatch board is bleeding profit margin
Most HVAC companies miss more calls than they answer during peak periods, based on our experience. Your dispatcher spends the day reacting instead of managing. Routes get inefficient. Technicians sit idle. And somehow, even as your revenue grows, your profit margins shrink.
ReadThe fleet costs bleeding your trades business dry
You track fuel. You budget for maintenance. You think you understand your fleet costs. But while you're watching the obvious expenses, the hidden ones are bleeding your business dry. Based on our experience across dozens of deployments, most trades businesses miss half their actual fleet costs.
ReadPeak season hiring destroys your profit margins
Every spring, you panic. Heat wave hits, calls flood in, and you realize you need more technicians. Now. So you scramble to post jobs, rush through interviews, and hire whoever shows up. By the time peak season ends, you've burned through budget on emergency wages, signing bonuses, and turnover costs.
Read
All articles16 articles