Harrison Chase on Deep Agents, LangSmith, and Earning Trust
Current thinking on trust and reliability, the two things that kill agent features.
Watch on YouTube youtube.com →
Building & Product
Video
Video
Podcast
Podcast
Podcast
Podcast
Podcast
Guide
Guide
Guide
Article
Article
Guide
Article
Article
Article
Article
Article
Should I put an AI agent in my product, or is a simple workflow enough?
A workflow follows steps you defined with the model filling in the reasoning slots; an agent decides its own path with tools in a loop. Workflows are cheaper, faster, and debuggable, and they cover most real product needs (tagging, drafting, summarising, extracting); agents earn their complexity only when the path genuinely cannot be known in advance. Both OpenAI and Anthropic say the same thing in their own guides: start with the simplest thing, add autonomy only when a measured need appears.
Go deeper
18 resources.
Everything Gets Rebuilt: The New AI Agent Stack
A stack-level tour of what you need around the model to make agents dependable.
Watch on YouTube youtube.com →Harrison Chase on Building the Orchestration Layer for AI Agents
The LangChain founder on what separates agent demos from agents in production.
Open sequoiacap.com →Transforming Customer Service through Company Agents, with Bret Taylor
The Sierra founder on where agents genuinely outperform, from thousands of deployments.
Listen on Spotify open.spotify.com →Make product management fun again with AI agents
Includes a five-question checklist for spotting low-risk, high-impact agent use cases.
Open lennysnewsletter.com →Vertical AI Agents | Bret Taylor (Sierra, OpenAI)
Why agents win when scoped to one vertical job rather than general autonomy.
Open pigment.com →Notion's Token Town: 5 Rebuilds, 100+ Tools
Notion rebuilt its agent architecture five times; hear what finally worked at 100M-user scale.
Open latent.space →Building Effective AI Agents
The most-cited engineering post on agents: five workflow patterns and when a true agent is warranted.
Open anthropic.com →A practical guide to building agents
OpenAI's distilled deployment lessons: use-case selection, guardrails, and orchestration patterns.
Open openai.com →12-Factor Agents: principles for production LLM software
The manifesto that successful agents are mostly well-engineered software with LLM sprinkled at key points.
Open github.com →A Developer's Guide to Building Scalable AI: Workflows vs Agents
A working engineer's decision criteria with cost and reliability tradeoffs spelled out.
Open towardsdatascience.com →AI Agents vs Workflows: When to Use Each
Short and sharp: if the steps are known in advance, a workflow wins on every axis.
Open redis.io →AI Workflows vs AI Agents
A neutral reference explainer to align your team on terminology before the build debate.
Open promptingguide.ai →AI agents vs. predefined workflows: practical decision guide
How a regulated enterprise makes the call, useful discipline for any founder.
Open medium.com →AI Agents vs. AI Workflows: Why Pipelines Dominate
Brings the production data: only 23% of orgs have scaled an agent anywhere.
Open intuitionlabs.ai →AI Workflows vs Agents: Why Workflows Win
The founder-flavoured argument against reaching for the harder, smarter-sounding option.
Open buildmvpfast.com →LangChain's CEO: better models alone won't get your agent to production
Counters the "just wait for GPT-next" excuse; the engineering around the model decides success.
Open venturebeat.com →Ambient Agents and the New Agent Inbox
The next interaction pattern: agents that act on events, with humans approving at checkpoints.
Open inferencebysequoia.substack.com →More in AI in Your Product
- Should my startup add AI features, and which ones actually matter to users? 20 →
- How do founders ship their first AI feature (APIs, wrappers, RAG in plain terms)? 19 →
- How much does it cost to run AI features, and how do I keep costs sane? 19 →
- Which AI model should I build on (GPT, Claude, Gemini, open-source), and how do I avoid locking myself in? 17 →
- How do I test whether my AI feature actually works before launch (evals, in plain terms)? 18 →
- How do I stop my AI feature from making things up in front of customers? 17 →
- Prompt engineering, RAG, or fine-tuning: which one does my product actually need? 18 →
- How do I design AI features people trust (chat vs buttons, streaming, showing the work)? 18 →
- What is my moat if I'm building on someone else's model? 19 →
- Is it safe to send customer data to AI APIs, and what about privacy laws like India's DPDP Act? 18 →
- How should I price AI features: per seat, per use, or per outcome? 17 →
- My AI feature is live. How do I monitor it and make it better with real usage? 18 →
Also in Starting Up
The same ground, over in Build the product, our Starting Up track.
eChai Partner Brands
