# Agent Report

> Source-based reporting and analysis on AI agents, published by Arach Tchoupani.

Published content only. Cite canonical article URLs and preserve attribution, uncertainty and disclosures.

## Navigation

- [Reports](https://agentreport.io/reports/)
- [Topics](https://agentreport.io/topics/)
- [Search](https://agentreport.io/search/)
- [Our method](https://agentreport.io/how-we-test/)
- [About](https://agentreport.io/about/)
- [Agent reading guide](https://agentreport.io/agents.md): Search and content formats
- [Site manifest](https://agentreport.io/agents.json): Versioned navigation and endpoints
- [Search API](https://agentreport.io/openapi.json): Full-text search and filters

## Reports

- [AI receptionist comparison: 46 offers by job, price and handoff](https://agentreport.io/articles/ai-receptionist-market-comparison.md): Compare 46 receptionist products and regional offers by documented workflows, plan conditions, human fallback and billing units. Includes source-linked scores and downloadable data.
- [DigitalOcean turns agent trouble into a test case](https://agentreport.io/articles/digitalocean-signals-agent-regression-tests.md): Signals and Simulations enter public preview, connecting production traces to candidate tests. Supported harnesses and mocked actions define the limits.
- [Meta brings Muse Wonder to developers, with memory at the center](https://agentreport.io/articles/meta-muse-wonder-agent-memory.md): Wonder reaches the API and Muse Code; Full requires a Power or Maximum subscription. The broader Muse rollout is still ahead.
- [Sakana's agents learn from their own team, with uneven gains](https://agentreport.io/articles/sakana-mass-self-supervision-limits.md): MASS reports better research scores per output token after two training cycles. Software benchmarks expose the limits of that improvement.
- [AWS Strands Box makes agent permissions depend on earlier actions](https://agentreport.io/articles/aws-strands-box-agent-sandbox.md): The sandbox preview can apply history-aware rules to agent tools. Direct file access sits outside that policy history, a boundary builders need to understand.
- [Cloudflare Clef-omni brings audio and video to agent decisions](https://agentreport.io/articles/cloudflare-clef-omni-agent-decisions.md): The new model scores permitted choices from mixed media. Its documented context limits and uneven benchmark results matter before those choices trigger actions.
- [Goodfire and Baseten bring live agent monitoring to open models](https://agentreport.io/articles/goodfire-agent-monitoring.md): Project Beacon aims to flag risky behavior during execution. The buying question is whether earlier detection produces an alert your team can act on.
- [From spreadsheet to invoice draft: where an AI agent should stop](https://agentreport.io/articles/ai-invoice-drafts-spreadsheet.md): Test invoice preparation with sample job records, duplicate checks and a human approval step before connecting an agent to real billing.
- [Is an AI receptionist worth it for your small business?](https://agentreport.io/articles/ai-receptionist-small-business-trial.md): Compare recovered work with the full bill and supervision time. Use a downloadable checklist to test the paid plan, bookings, human handoffs and rollback.
- [Use AI in a shared inbox without losing track of who replies](https://agentreport.io/articles/ai-shared-inbox-human-approval.md): A practical first workflow for small teams: assign ownership, prepare AI drafts, check for new messages and test handoffs before allowing automatic sends.
- [Who keeps your business agent working after setup?](https://agentreport.io/articles/maintain-small-business-ai-agent.md): Give a small-business agent a named owner, backup, change log and pause procedure. Includes a handover card and maintenance test sheet.
- [Is your AI agent saving time? A small-business pilot worksheet](https://agentreport.io/articles/measure-ai-agent-roi-small-business.md): Count setup, checking, repairs and software costs before calling an AI pilot a success. Includes a blank ledger and a worked example.
- [Agents are crossing app boundaries. The handoff matters.](https://agentreport.io/articles/cross-app-agents.md): A source-based briefing on Google Workspace and Claude: where tasks start, where work happens, and what users still need to verify.
- [An agent should leave you with proof](https://agentreport.io/articles/how-we-test.md): Our testing method: a real task, explicit permissions, a result you can inspect, and a record of the work still left to you.
- [What is Instinct? The personal assistant behind the name](https://agentreport.io/articles/instinct-personal-assistant.md): What Instinct says it can do, who is behind it, and the questions a real-world trial needs to answer.

## Optional

- [Full catalogue](https://agentreport.io/articles.json)
- [Complete edition](https://agentreport.io/llms-full.txt)
- [RSS](https://agentreport.io/feed.xml)
- [JSON Feed](https://agentreport.io/feed.json)