SemiAnalysis found Perplexity's enterprise Slack integration 'shocking better than Claude' in internal tests. The AI research firm tried Devin, Mistral, Grok, and Cerebras alongside standard tools.
Key facts
- 96% of SemiAnalysis token budget goes to Anthropic.
- Perplexity Slack integration called 'shocking better than Claude'.
- Tools tested: Devin, Mistral, Grok, Perplexity, Cerebras, Claude, Codex.
- Most tools churned as 'negative value add'.
- Perplexity bot sends images and returns reports in Slack.
SemiAnalysis, the AI research and analysis firm, this week conducted an internal bake-off of multiple AI coding and research tools. According to a post by @SemiAnalysis_, the company tested Devin, Mistral, Grok, Perplexity, Cerebras Public API, and the usual Claude and ChatGPT/Codex tools. Most were dismissed quickly as having 'negative value add'.
One tool stood out: Perplexity's enterprise Slack integration. The firm described it as 'shocking better than Claude,' specifically for its ability to send images and return detailed research reports and visualizations directly in Slack. The integration, led by Perplexity CEO Aravind Srinivas, is a Slack-native bot that ingests user queries with images and outputs structured analysis.
The results challenge Anthropic's dominance in enterprise AI. SemiAnalysis currently allocates 96% of its token budget to Anthropic models, but based on usage patterns, the firm suggests some spend may shift to Perplexity. The post explicitly calls on Anthropic product lead Cat Wu to add similar features to Claude's Slack integration.
The unique take: This is not about raw model quality — it's about workflow integration. Perplexity's Slack bot wins on UX and output format (visualizations, structured reports) rather than pure reasoning. For enterprise AI adoption, integration fidelity may matter more than benchmark scores.
What this means
Perplexity's enterprise play is increasingly direct. The company has been building Slack and enterprise features since late 2025, but this is the first public comparison where a sophisticated AI team (SemiAnalysis) calls it better than Claude for a specific use case. The churn of other tools — Devin, Mistral, Grok — suggests the bar for enterprise AI tooling is high: negative value add is common when tools don't fit existing workflows.
Anthropic has been investing heavily in enterprise features, including Claude Pro and enterprise Slack integrations, but Perplexity's advantage may be in responsiveness and output format. The SemiAnalysis team wants Claude to add image-in, report-out capabilities to its Slack integration — a feature Perplexity already ships.
Key Takeaways
- SemiAnalysis found Perplexity's Slack bot beats Claude in internal trial.
- 96% token budget goes to Anthropic, but usage may shift.
What to watch
Watch for Anthropic's response: if Cat Wu adds image-in, report-out features to Claude's Slack integration in the next quarter, it signals competitive pressure. Also track Perplexity enterprise ARR growth — SemiAnalysis is a proxy for sophisticated buyers.
