IAAI-27 Timetable for Authors

All deadlines are anywhere on earth (11:59 PM UTC-12) unless otherwise noted.

July 28, 2026: 
IAAI-27 OpenReview submission site opens for paper submission

September 8, 2026: 
Electronic paper submissions due

October 30, 2026: 
Notification of decisions (no rebuttal period)

November 24, 2026: 
Final decisions on conditionally accepted papers

December 14, 2026:
Camera-ready copy due

February 18 to 20, 2027:
IAAI-27, co-located with AAAI-27, Montreal, Canada

IAAI-27 Call

The Thirty-Ninth Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-27) is the premier venue for studies of innovative, impactful AI applications in the real world and for insights into the process of developing, deploying, and maintaining them. IAAI-27 will feature technical papers, best practices, invited talks, and panel discussions that explore challenges, methodologies, and lessons learned from the practical use of AI.

A strong IAAI paper answers questions like the following. What new domain or problem was explored, and why was it interesting or hard? What new technical approach was applied, and what design decisions and alternatives shaped the solution? What impact did the application have, and how was it measured? What happened on the way to deployment: what went wrong, what was redesigned, and what was learned that others can use? 

Submissions should focus on applied AI; purely theoretical work and algorithmic descriptions are more suited for AAAI-27. The exception: papers in which a new technical approach solved a real-world problem that prior approaches could not. Such papers must show why prior approaches fell short and how the new development overcame those limits, with the emphasis on the deployed solution and its impact. To demonstrate real-world impact, co-authors from the deploying organization are recommended. 

Topics of special interest for IAAI-27 include but are not limited to: deployed applications of large language models and foundation models, including agentic systems in production; evaluation and assurance of AI systems in deployment; the economics of operating AI at scale; optimization, planning, and scheduling in live operations; classical machine learning in enterprise production; knowledge-based, symbolic, and hybrid applications; applications operating under regulatory regimes such as the EU AI Act; embodied and physical AI; and AI applied to science and engineering.

IAAI-27 seeks contributions in the following areas: (1) Deployed Applications showcasing the use of AI with measurable benefits, best practices, and lessons learned; (2) Emerging Applications of AI on trajectory for full deployment; (3) Deployment Insights: tools, practices, evaluations, and candid accounts of what went wrong and what was learned in building and operating AI systems, submitted either as full papers or, new for IAAI-27, as short Experience Reports written by practitioners themselves.

Tracks and Topics 

1. Deployed Highly Innovative and Impactful Applications of AI

Papers submitted to this track must describe deployed applications that demonstrate measurable benefits through innovative use of AI technology. An application is considered deployed once it is in production and used by end-users, with meaningful data collected on its performance. Submissions may focus on either standalone applications or components of larger systems. Innovative applications built on top of commercial or open foundation models are fully in scope; the innovation may lie in adaptation, integration, evaluation, or operation rather than in the underlying model.

Papers will be evaluated based on the quality of the problem description, AI approach, innovation in AI use, design rationale (the design decisions made and the alternatives considered), measurable improvements, deployment details, and lessons learned during development, deployment, and maintenance, where accounts that report failures, setbacks, and abandoned approaches alongside successes will be evaluated more favorably than success-only narratives. Particular weight is also given to the novelty of the application domain or problem: papers that open a new domain for AI, or that solve a problem previously considered impractical, will be weighted accordingly. In-depth algorithm descriptions are better suited for AAAI. Each accepted paper in this track will receive the IAAI Deployed Application Award.

Page Limit: 8 pages. Format: AAAI style and formatting guidelines. No page limit for references and appendices.

2. Emerging Applications of AI 

The Emerging Applications track focuses on applications of AI methods to real-world problems that are not yet fully deployed but include early deployment or pilot-stage results. Submissions must demonstrate a clear path toward full-scale deployment, addressing emerging engineering or sociotechnical challenges and their practical relevance. Papers will be evaluated on problem significance, innovation, AI methodology, design rationale, technical quality, and clarity, with a clear path toward deployment. The quality threshold for acceptance rises with distance from deployment: a paper reporting pilot results with real users faces the standard bar, while a paper whose deployment lies further ahead must compensate with stronger evidence and analysis. Papers without an articulated path to deployment may be rejected without review of the paper substance.

Page Limit: 6 pages. Format: AAAI style and formatting guidelines. No page limit for references and appendices.

3. Deployment Insights: Tools, Practices, and Lessons from the Field 

This track collects the knowledge that makes the applied AI community collectively better at building and operating AI systems: deployed tools and methodologies, evaluation and assurance practices, and honest accounts of failures and their remediation, across all AI technologies.

Areas of interest include, but are not limited to:

Path to Responsible Deployment: evaluation harnesses, regression testing across model, data, and rule-base versions, red-teaming practices, and pre-deployment safety cases in production use, including validation of automated evaluation approaches such as LLM-as-judge.

Operating AI Systems in Production: monitoring, observability, guardrails, rollback, and human-escalation mechanisms for AI systems of any kind, including the tracing and sandboxing of agentic and generative systems.

What Went Wrong and Why: analysis of failures, setbacks, and incidents experienced in real deployments of any AI technology, from data drift and model degradation to failure modes specific to generative and agentic systems such as prompt injection, hallucination-driven harms, and unintended agent actions; analysis of trends across such failures; and practices that prevented or mitigated harm.

Deployment Economics and Efficiency: cost and latency engineering with measured production impact, from classical resource optimization to model routing, caching, quantization, and distillation for large-model serving.

Deployed Data Quality Tools: tools in use that address challenges in processing raw data in AI systems, from sensor and enterprise data pipelines to training-data curation and retrieval-corpus quality.

Compliance and Documentation in Practice: audit trails, model documentation pipelines, and conformity assessment practices driven by regulatory requirements.

Impact on Business Models: accounts of how deployed AI changed what an organization sells, how it prices, or how it operates, including new AI-enabled products and services, shifts from labor-based to outcome-based delivery, build-versus-buy decisions and their consequences, and cases where an AI deployment forced or failed to force a change in the business model itself.

Contributions to this track may be submitted in one of two formats:

3a. Full Papers

Papers will be evaluated primarily by the following criteria: the extent to which the work advances real-world application quality, development, deployment, maintenance costs, productivity, reliability, and/or scaling, and the degree to which it enables applied AI practitioners to produce better AI solutions, including ease of use, adoption, and documentation. Tools that are deployed will be ranked more highly in the review process. Analysis of failures must center on failures experienced during AI deployments, with systems directly impacting the real world. Tools and methodologies still in the concept phase, or tested only in laboratory settings, are not suitable for this track.

Page Limit: 6 pages. Format: AAAI style and formatting guidelines. No page limit for references and appendices.

3b. Experience Reports (new for IAAI-27)

Focused, first-person accounts from practitioners of a deployment experience, a failure and its remediation, or an operational practice, in any of the areas above, without the framing burden of a full technical paper. Experience Reports are reviewed by the same committee under the same single-blind process, with evaluation adjusted for length and weighted toward the usefulness and candor of the insight. Authors from the deploying organization are expected.

Page Limit: 2 to 4 pages. Format: AAAI style and formatting guidelines. No page limit for references and appendices.

Submissions in both formats will be reviewed by a program committee of AI researchers, industry experts, human factors, safety, and ethics specialists, with a particular focus on whether the paper efficiently informs the production and deployment of AI systems.

Submission Instructions

Electronic submissions are required. Papers must be in trouble-free, high-resolution PDF format, US letter (8.5″ x 11″) and AAAI two-column format (use the CameraReady templates in the AAAI-27 Author Kit). Please see track descriptions for specific page limits above.

Supplementary material in the form of (1) technical appendix, (2) multimedia, (3) code, and (4) data may be submitted alongside all papers. The main submission may reference the supplementary material, but it should be self-contained. Evaluations are primarily based on the main submission, and reviewers are not obliged to consult the supplementary material.

Submissions to IAAI-27 should contain identifying information about the authors and their affiliations. Reviewing is single-blind (anonymous reviewers), NOT double-blind. It is also recommended that authors from the deploying institutions be involved as co-authors to show the real-world impact of the submission.

Authors should register on the IAAI-27 OpenReview paper submission site. IAAI does not accept papers submitted through other channels. Submissions received after the deadline or not meeting the length or formatting requirements detailed above will not be accepted for review. Notification of the electronic paper receipt will be emailed to the first author (or designated contact author) soon after receipt.

By submitting a paper to IAAI, the authors agree that the Chairs have the final decision on the acceptance of publication at the conference. Authors will not be invited to rebut their reviews. Some acceptances may be conditional on specific revisions, completed in coordination with the Program Committee ahead of the camera-ready deadline.

All submissions must be original. IAAI-27 will not consider any paper that, at the time of submission, is identical or substantially similar to another paper that is submitted concurrently to IAAI-27, or that is under review, accepted for publication, already published or accepted for publication in a journal or another conference. Once submitted to IAAI-27, authors may not submit the paper or a substantially similar version elsewhere during IAAI’s review period. These restrictions apply only to refereed journals and conferences, not preliminary versions posted as preprints to arXiv.org or other unrefereed forums, or workshops with a limited audience and without archival proceedings. Authors must confirm that their submissions to IAAI conform to these requirements at the time of submission.

Important Dates

All deadlines are anywhere on earth (11:59 PM UTC-12) unless otherwise noted.

July 28, 2026: IAAI-27 OpenReview submission site opens

September 8, 2026: Electronic paper submissions due

October 30, 2026: Notification of decisions (no rebuttal period)

November 24, 2026: Final decisions on conditionally accepted papers

December 14, 2026: Camera-ready copy due

February 18 to 20, 2027: IAAI-27, co-located with AAAI-27, Montreal, Canada

---