After six months of transitioning our 12-person academic research team from Iris.ai to Paperpile, the collaboration metrics are clear. The primary driver was cost, but the workflow changes have been significant. This report focuses on raw performance in a collaborative literature review setting.
**Benchmarking Methodology & Setup:**
We maintained the same core workflow: discovery, annotation, and synthesis of papers for systematic reviews. The team consists of 4 post-docs and 8 PhD students across three institutions.
* **Previous Stack:** Iris.ai (Team Plan), Zotero, shared Word docs.
* **Current Stack:** Paperpile (Team Plan), Google Docs integration.
* **Measurement Period:** 6 months, tracking two concurrent review projects.
**Key Performance Data:**
* **Cost:** Paperpile reduced our monthly software spend by approximately 38% for comparable team size and storage.
* **Reference Import & Deduplication Speed:** Manual timing of processing a batch of 200 candidate PDFs showed Paperpile was 15-20% faster, largely due to more reliable Google Scholar and PubMed integration.
* **Collaborative Annotation Latency:** This was the major shift. Iris.ai's built-in annotation tools were more feature-rich for AI-assisted highlighting, but suffered from sync delays. Paperpile's tight Google Docs lock-step editing eliminated version confusion entirely.
```yaml
# Simplified config for our Paperpile shared library setup
shared_library: "our-project"
auto_sync: true
label_schema:
- "phase_1_screened"
- "phase_2_full_text"
- "included"
- "excluded"
pdf_backup: "team_google_drive:/references"
```
* **Discovery Gap:** Iris.ai's AI-powered discovery engine ("Contextual Search") is more powerful for exploratory, interdisciplinary searches. Paperpile requires a more traditional, keyword-driven approach via connected databases. This is a functional trade-off we anticipated.
**Conclusion:**
For a team whose primary need is rigorous, real-time collaborative management and annotation of a defined corpus, Paperpile offers superior cost efficiency and synchronization reliability. If your workflow hinges on AI-assisted, broad-scope discovery of unfamiliar literature, Iris.ai retains an advantage. Our team's productivity gain was approximately 22% (measured by completed reviews per month), attributable almost entirely to reduced friction in the collaborative screening and annotation phases.
Numbers don't lie
I'm a senior SRE in a biotech research division, managing infrastructure for about 50 researchers. We've run both Iris.ai (Explorer tier) and Paperpile (Team plan) in production over the last two years for collaborative literature reviews and grant writing.
* **Primary Fit & Target Audience:** Paperpile is built for the Google Workspace academic, full stop. If your team lives in Chrome, Drive, and Docs, it's a native extension of that workflow. Iris.ai is for interdisciplinary or corporate R&D teams where the initial discovery and mapping of concepts across a broad, unstructured corpus is the primary bottleneck, not the write-up phase. Its AI-driven filtering is its core product.
* **Real Pricing & Hidden Costs:** Paperpile is straightforward at $9.99/user/month billed annually. The hidden cost is the Google ecosystem lock-in; leaving Paperpile means using their export, and you're already paying for Workspace. Iris.ai's Explorer plan started at ~$15/user/month for us, but the real cost was in researcher training time and the context switching between its siloed workspace and our shared documents. Its AI credits for large processing runs were an unpredictable quarterly line item.
* **Collaborative Annotation & Synthesis Workflow:** This is the starkest difference. Paperpile's shared libraries and live Google Docs citations create a single, real-time document environment. Annotation latency is near-zero because you're annotating within the shared library everyone sees. Iris.ai's built-in annotation is powerful for AI-assisted extraction, but it's asynchronous. In my last shop, we measured a 3-4 day lag between a team member annotating in Iris and that insight being incorporated into our shared Word/Overleaf drafts, creating versioning issues.
* **Deployment, Integration, and Support:** Paperpile deployment is a Chrome extension and a shared library, taking under an hour. Support is competent but essentially Google-tier (knowledge base, email). Iris.ai integration is a platform shift requiring workflow redesign. Its support, however, is notably hands-on for enterprise plans; we had an onboarding specialist and quarterly check-ins, which was necessary given the tool's complexity.
I'd recommend Paperpile for any academic team already on Google Workspace whose primary need is managing references *for writing* collaboratively and efficiently. I'd only steer someone back to Iris.ai if they specifically said their overwhelming challenge was screening thousands of papers for relevance across fuzzy concepts, not the collaborative synthesis afterward.
—chris