Collabagents Taɣerɣert n Umɛiwen s Waṭas n Yimeggiyen sɣur Jiahui Yu and Xudong Xing

Collabagents Taɣerɣert n Umɛiwen s Waṭas n Yimeggiyen sɣur Jiahui Yu and Xudong Xing d anerbaz yettwaqadren meqqren n 2026 Tarbaḥt A' Design Award n Lbrunz deg Taggayt n Tnaflit n Afeṣṣel n Ugrudem, Amyigaw d Tarmit n Useqdac.

LbrunzAgrudem· 2026

Collabagents Taɣerɣert n Umɛiwen s Waṭas n Yimeggiyen

sɣur Jiahui Yu and Xudong Xing

Amsaɣ: CollabAgents

Anerbaz n Tarbaḥt A' Design Award n Lbrunz deg Taggayt n Agrudem

9,696

Anerbaz n Urraz Yettwasenqden

Senqed Addad n Urraz i #183243

Wali deg A' Design Award Tabḍert n tɣamsa Tanecḍa n wammud

Initen

Afeṣṣel

Ɣef Collabagents Taɣerɣert n Umɛiwen s Waṭas n Yimeggiyen

CollabAgents d taɣerɣert n umɛiwen s waṭas n yimeggiyen yettwaxedmen i yiselkimen n umahil uddizen. Iseqdacen snulfuyen-d aselkim n umahil, sefkan-d timlilin i yemdanen yettekkan deg-s. Mi ara texṣeṛ tamlilt, yettawi-tt umeggi n tussna tuhlisant iwakken ad ikemmel ubrid n umahil. Amahil yettwaseggem d tiqḍiwin yeqqnen gar-asent, ma yella yettwabeddel yiwet n teqḍimt, ad ttubeddlen ula d yisurifen i d-iteddun deffir-es. Taywalt tuddizt yebnan ɣef tengalt tettwaseqdac iwakken ad tettwazen telɣut s tseddi, ad tenqes tirit. CollabAgents yettarra iselkimen n umahil imezgiyen d yinagrawen imyifen i d-yettarran awal ɣef ubeddel.

Asmekti

Ɣef Tarbaḥt A' Design Award n Lbrunz

Collabagents Taɣerɣert n Umɛiwen s Waṭas n Yimeggiyen d amerbaḥ n Tarbaḥt A' Design Award n Lbrunz, tagmat ɛlayen i yisnulfa yelhan, yeṭṭafaren tinawt, tamusni d tezmert. Imerbaḥen n Lbrunz d isnulfa yelhan, ttbanen d timusniwin n tmusni n yimahilen-nsen. Collabagents Taɣerɣert n Umɛiwen s Waṭas n Yimeggiyen ad yili d asnulfu yettwaɛezzen, yelhan, amaynut. Collabagents Taɣerɣert n Umɛiwen s Waṭas n Yimeggiyen yettemlil ɣef yimawlan, yettwaɛezzen s tmahilt-is.

Deffir n ufeṣṣel

Tagrurt n tɣamsa

Ismawen n terbaɛt, iferdisen itiknikanen akked telɣut i d-yefka umeṣṣufeɣ.

Taɣara Tasuft: CollabAgents is a multi-agent collaboration system that simulates missing roles with AI agents to ensure continuous workflows. Node-level editing enables targeted updates, reducing rework by over 85%, while code-based communication improves precision and speed. It transforms linear workflows into adaptive, resilient systems.
Tigejdiwin & Akemmus: The system features a node-based workflow structure, role simulation modules, and dynamic process propagation. It supports real-time updates, modular task execution, and scalable multi-agent coordination. Key capabilities include node-level editing, automatic downstream updates, and role auto-completion to maintain continuous workflow execution.
Tibzimin: multi-agent, collaboration, workflow, AI system, interaction design, role simulation, automation, UX, decision system
Tasnakta n Ufares: CollabAgents is built using AI-driven multi-agent architecture, workflow modeling, and structured data exchange. The system leverages role-based agent simulation, node-level state management, and real-time synchronization to enable adaptive collaboration. Code-based communication ensures precise information transfer and reduces ambiguity across processes.
Asewḥel n Usnulfu: The main challenge was designing a system that balances automation with user control in complex workflows.

Traditional collaboration relies heavily on human coordination, making it difficult to handle missing roles, dynamic changes, and process interruptions. The challenge was to create a system that allows targeted intervention without breaking the overall workflow.

This was addressed through node-level editing and role simulation, enabling flexible adjustments while maintaining system stability and continuity.
Anadi: The project is based on research into cross-functional collaboration challenges, including workflow inefficiencies, role dependency, and communication breakdowns.

Methods included workflow analysis, scenario mapping, and observation of real-world team processes. The research identified key issues such as delays caused by missing roles and the need for more adaptive systems.

The design explores how multi-agent systems can simulate roles, improve coordination, and reduce rework, resulting in more efficient and resilient collaboration.
Aseḥbiber: CollabAgents is inspired by the complexity of cross-functional collaboration, where delays often result from missing roles and fragmented communication. It explores how AI agents can simulate expertise and fill capability gaps in real time, transforming collaboration into a more adaptive and efficient process.
Tanzagt n Usenfar: The project was developed between 2025 and 2026 in the United States as an independent concept design focused on future collaboration systems.
Tanila n Umahil: Users start by selecting or creating a workflow and assigning roles. If roles are missing, AI agents automatically step in to continue the process.

Users can interact with specific steps, make changes, and instantly see updates applied to the rest of the workflow. Instead of restarting the entire process, only affected parts are adjusted.

This enables smoother collaboration, faster decision-making, and reduces delays caused by miscommunication or missing inputs.
Ibrivan: No patents or trademarks are registered for this project.
Igmamen n Terbaɛt: Jiahui Yu
Xudong Xing
Imsuren n Tugniwin: Image #1: Creators Jiahui Yu and Xudong Xing, CollabAgents, 2025.
Image #2: Creators Jiahui Yu and Xudong Xing, CollabAgents, 2025.
Image #3: Creators Jiahui Yu and Xudong Xing, CollabAgents, 2025.
Image #4: Creators Jiahui Yu and Xudong Xing, CollabAgents, 2025.
Image #5: Creators Jiahui Yu and Xudong Xing, CollabAgents, 2025.