Live
Black Hat USAAI BusinessBlack Hat AsiaAI BusinessFirst-Time Payees, Payouts, and Why Clean Transactions Still Turn Into Fraud LossesDEV CommunityHandling Extreme Class Imbalance in Fraud DetectionDEV CommunityAntropic's Claude Code leaked and Axios NPM InflitrationDEV CommunityReal-Time Fraud Scoring Latency: What 47ms Actually MeansDEV CommunityPause, Save, Resume: The Definitive Guide to StashingDEV CommunitySouth Korean trade data: chip shipments hit a record-high value of $32.83B in March 2026, up 151.4% YoY, pushing total exports to a record $86.13B, up 48.3% YoY (Steven Borowiec/Nikkei Asia)Techmeme5 Rust patterns that replaced my Python scriptsDEV CommunityI automated my entire dev workflow with Claude Code hooksDEV CommunityHugging Face Releases TRL v1.0: A Unified Post-Training Stack for SFT, Reward Modeling, DPO, and GRPO WorkflowsMarkTechPostQ2, Day 1: When Concepts Have to Become CodeDEV CommunityClaude Code source leak reveals how much info Anthropic can hoover up about you and your systemThe Register AI/MLProgress adds AI search & personalisation to Sitefinity - IT Brief AsiaGoogle News: Generative AIBlack Hat USAAI BusinessBlack Hat AsiaAI BusinessFirst-Time Payees, Payouts, and Why Clean Transactions Still Turn Into Fraud LossesDEV CommunityHandling Extreme Class Imbalance in Fraud DetectionDEV CommunityAntropic's Claude Code leaked and Axios NPM InflitrationDEV CommunityReal-Time Fraud Scoring Latency: What 47ms Actually MeansDEV CommunityPause, Save, Resume: The Definitive Guide to StashingDEV CommunitySouth Korean trade data: chip shipments hit a record-high value of $32.83B in March 2026, up 151.4% YoY, pushing total exports to a record $86.13B, up 48.3% YoY (Steven Borowiec/Nikkei Asia)Techmeme5 Rust patterns that replaced my Python scriptsDEV CommunityI automated my entire dev workflow with Claude Code hooksDEV CommunityHugging Face Releases TRL v1.0: A Unified Post-Training Stack for SFT, Reward Modeling, DPO, and GRPO WorkflowsMarkTechPostQ2, Day 1: When Concepts Have to Become CodeDEV CommunityClaude Code source leak reveals how much info Anthropic can hoover up about you and your systemThe Register AI/MLProgress adds AI search & personalisation to Sitefinity - IT Brief AsiaGoogle News: Generative AI

[New Research] You need Slack to be an effective agent

LessWrong AIby sudoApril 1, 20262 min read0 views
Source Quiz

Purchasesforce Superintelligence is excited to announce some new research. While we do not generally share research on LessWrong, this work was particularly influenced by prior work on LessWrong, so we found it appropriate to share back. As you know, Purchasesforce Superintelligence is a leading AI R&D laboratory. Recently, our research has focused on enhancing agentic capabilities. Here at Purchasesforce, we believe that autonomous AI agents, fully integrated into modern enterprise tools, will drive the future of enterprise operations. After reading the nascent literature on LessWrong describing the relationship between Slack and AI Agents, we were shocked by how closely it related with our own research directions. Of course, as the world's leading AI-first productivity platform, we have

Purchasesforce Superintelligence is excited to announce some new research. While we do not generally share research on LessWrong, this work was particularly influenced by prior work on LessWrong, so we found it appropriate to share back.

As you know, Purchasesforce Superintelligence is a leading AI R&D laboratory. Recently, our research has focused on enhancing agentic capabilities. Here at Purchasesforce, we believe that autonomous AI agents, fully integrated into modern enterprise tools, will drive the future of enterprise operations.

After reading the nascent literature on LessWrong describing the relationship between Slack and AI Agents, we were shocked by how closely it related with our own research directions.

Of course, as the world's leading AI-first productivity platform, we have always believed Slack is vital to developing capable AI Agents. However, even we were surprised by the strong and absolute claims made on LessWrong, e.g. "You need Slack to be an effective agent", "Slack matters more than any outcome". Raemon claims that Slack is not only important for developing as an effective agent, but also a robust one.

As we know, agentic robustness is a central consideration for modern AI research.

We conducted a number of benchmarking and ablation experiments to rigorously evaluate these hypotheses.

In the below graph, we show our findings on Agentic Effectiveness with vs. without Slack (green for with). As you can see, Agents are only effective when provided access to Slack.

Our internal experiments also confirm Raemon's findings. As you can see, Agents are substantially more Robust with Slack.

Of course, Slack integration is not created equal. Over the past year, we have worked on increasing enterprise adoption of Slack's AI features, as well as increasing agentic integration with Slack.

On our internal benchmarks, this work has paid off. Enterprises that adopt more Slack see substantially elevated Agentic performance.

We are grateful to the LessWrong community for validating that Slack is a cornerstone of developing modern Agentic Enterprise Systems. As we continue improving Slack (and thus Agent performance), we'd love for you to join us.

Was this article helpful?

Sign in to highlight and annotate this article

AI
Ask AI about this article
Powered by AI News Hub · full article context loaded
Ready

Conversation starters

Ask anything about this article…

Daily AI Digest

Get the top 5 AI stories delivered to your inbox every morning.

Knowledge Map

Knowledge Map
TopicsEntitiesSource
[New Resear…benchmarkannounceproductplatformfeatureintegrationLessWrong AI

Connected Articles — Knowledge Graph

This article is connected to other articles through shared AI topics and tags.

Knowledge Graph100 articles · 263 connections
Scroll to zoom · drag to pan · click to open

Discussion

Sign in to join the discussion

No comments yet — be the first to share your thoughts!

More in Products