
How do you get AI agents into day-to-day work? Tonight in NYC: @mitch_troy and Nikhil Lohchab on agent deployment with @btv_vc and Coast. 6–8:30pm ET. Request an invite → luma.com/9dyz6jv3
✓AI agents for accountants, at firms and in-house. Tax, audit, and CAS for firms. Reporting, forecasting, the close, and analysis for finance teams.

How do you get AI agents into day-to-day work? Tonight in NYC: @mitch_troy and Nikhil Lohchab on agent deployment with @btv_vc and Coast. 6–8:30pm ET. Request an invite → luma.com/9dyz6jv3

Long-horizon agents building long-horizon agents. We use Devin to build and test our accounting agents. One thread on our latest tax agent ran 20 days. Thanks @cognition for the feature.

"It's going to sound silly, but I promise it's very effective." @mitch_troy's favorite debugging trick: have a subagent pretend to be your prod agent. Tell it not to cheat. Replay the scenario and tweak the context. "You can do this today without building anything." @aiDotEngineer NYC:

AI made engineers much faster. Most companies didn't follow. @mitch_troy at @aiDotEngineer NYC on why: agents spread when they're simple and reliable, and reliability comes from opinions. Plus behavior specs over outcome evals, and debugging bad calls with forking and sub-agent simulations. 19 min: getbasis.ai/blogs/why-ai-mak…

"It failed the outcome. How do you know why? You have potentially a thousand inference calls in your 5 hour long agent trajectory." @mitch_troy on forking: rewind to right before the bad decision, fork it, and ask the agent why it did that. From @aiDotEngineer NYC:

Introducing Basis in your binder. Basis now runs inside the audit binder your firm already uses. Agents work right alongside your team, and auditors start at review. No moving documents between tools. No writing prompts. Your firm's standards apply on every engagement. Firms that start now head into busy season with agents already doing the first pass. Read the full post → getbasis.ai/blogs/introducin…

"What you don't want is a customer getting one way one time and another way the other time." @mitch_troy on behavior specs: "a product spec for the agent era." Write down the opinions that matter most, then use them as a rubric, even with no ground truth. @aiDotEngineer NYC:

"Agents as you think of them today outsource the complexity to the user." @mitch_troy on the two gaps blocking agents in production: competency and process. The fix is simple and reliable. "Nobody is looking for perfect." From @aiDotEngineer NYC.

"If you just said agent extract Excel, extract PDF, you might get good outcomes, but they're not going to be reliable." @mitch_troy: reliability comes from opinions, and some come from expertise. Read the primary source before answering a tax question. @aiDotEngineer NYC:

Electra uses Devin to write code. She also decides what needs doing, tests changes, ships updates to internal tools, and pushes the team. @matthew_busel shares what it takes to move from agents to actually delegating responsibility to AI teammates:

"Nobody wants their agent to move 37 their tax return." @mitch_troy: review agent work like a PR. Blast radius, and how much you trust the author. "For agents to diffuse, they need to be simple. For teams to trust them, they need to be reliable." From @aiDotEngineer NYC.

We’ve had a blast learning with you, and now we’re scaling up every piece of what we do. With the support of @trybasis, we’re going all in on our mission to become NYC’s home for designers and builders. In this next chapter, we’ll be focused on producing industry-shaping events and media for our community here in NYC and our friends beyond. Our own @falcone_design talked with @simoncorry and @mitch_troy about Basis betting on Double Diamond, and the power of new media and real community. (00:00) John and Simon join Basis to build Labs (00:34) The origin of Double Diamond (04:06) Why Basis is supporting Double Diamond (07:55) Why traditional recruitment is dead (11:46) Becoming a meat proxy (16:13) Babysitting Claude vs becoming prolific (25:21) Leading users to the future (31:34) Design systems are not solved (37:46) Nobody cares about accounting (42:12) Accounting as the foundation of capitalism

Excited for this come join to chat about the real work that happens when deploying long horizon and proactive agents to the real economy

Read about how we build agents with @cursor_ai !

Every person at Basis, no matter what team they're on, has agents doing work alongside them within minutes of their first day. We’ve been early to integrate AI into every process in our company - in part because we invested in building a team dedicated to making this happen. If that sounds interesting, you should come join the Atlas team! Read more from OpenAI about how we’re using Codex to make onboarding teachable with agents: openai.com/index/ai-native-c…

Last week, @mitch_troy joined @AI_in_the_AM to discuss accounting as “an intelligence over the economy,” how AI can help firms create capacity for growth, and how Behavior Specs can help supervise long-running agents. Watch the full conversation: