--- title: Work in progress (WIP) is the new software delivery... description: Coding agents let teams open far more branches at once, but median throughput on main hasn't improved. CircleCI's 2026 State of Software Delivery Q3 Pulse... canonical: https://daily.dev/posts/work-in-progress-wip-is-the-new-software-delivery-bottleneck-7xif44njt twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Work in progress (WIP) is the new software delivery bottleneck | daily.dev og:description: Coding agents let teams open far more branches at once, but median throughput on main hasn't improved. CircleCI's 2026 State of Software Delivery Q3 Pulse... og:url: https://daily.dev/posts/work-in-progress-wip-is-the-new-software-delivery-bottleneck-7xif44njt og:image: https://api.daily.dev/og/posts/7XIF44nJT.png og:image:alt: Work in progress (WIP) is the new software delivery bottleneck og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Work in progress (WIP) is the new software delivery bottleneck **[CircleCI](https://daily.dev/sources/circle)** · 8 min read · 1 upvotes · 0 comments ## Summary Coding agents let teams open far more branches at once, but median throughput on main hasn't improved. CircleCI's 2026 State of Software Delivery Q3 Pulse report finds active projects up 16% and active branches per contributor rising from 7 to 9, while median main-branch throughput stays flat at 1.7 runs per day. By Little's Law, doubling work in progress without raising completion rate doubles cycle time. The top 5% of organizations grew branches 55% but also raised main-branch throughput 31%. CircleCI recommends tracking WIP and cycle time, giving agents direct CI feedback, and measuring delivery rather than activity. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### why isn't shipping speed increasing even though my team uses more coding agents and opens more branches Work in progress has become the bottleneck, not coding capacity. CircleCI's Q3 2026 data shows active feature branches per contributor rose from 7 to 9 and active projects grew 16%, yet median main-branch throughput stayed flat at 1.7 runs per day, since validation and review capacity didn't grow at the same rate as parallel work. _Teams debugging stalled delivery despite more AI coding agents can track these same bottleneck patterns via daily.dev._ ### how does Little's Law explain why more parallel development with coding agents slows down delivery Little's Law states average cycle time equals average work in progress divided by average throughput. If agents double the number of changes in progress from 10 to 20 while merges per day stay at 5, cycle time doubles from 2 to 4 days; only matching the completion rate to the new WIP level keeps cycle time stable. _Engineers reasoning about agent-driven workflow capacity can dig deeper into these delivery models on daily.dev._ ### what metrics distinguish high-throughput engineering teams from median teams when scaling coding agents The top 5% of organizations by main-branch activity grew active feature branches by 55% while also increasing main-branch throughput by 31%, widening their lead over the median team's flat throughput to more than 10 times by Q3 2026. Key tracked metrics include WIP, cycle time, time to first review, and merge efficiency ratio (CI runs on branches relative to main). _Teams benchmarking delivery velocity against peers can follow this kind of throughput data on daily.dev._ ## Similar posts on daily.dev - [5 key takeaways from the State of Software Delivery Q3 Pulse report](https://daily.dev/posts/5-key-takeaways-from-the-state-of-software-delivery-q3-pulse-report-muizbv3x1) · CircleCI · 0 upvotes · 0 comments - [5 key takeaways from the 2026 State of Software Delivery: Why AI isn’t shipping faster](https://daily.dev/posts/5-key-takeaways-from-the-2026-state-of-software-delivery-why-ai-isn-t-shipping-faster-oktnhjrhb) · CircleCI · 4 upvotes · 0 comments - [What 986 million code pushes say about the developer workflow in 2025](https://daily.dev/posts/what-986-million-code-pushes-say-about-the-developer-workflow-in-2025-4bpyjwlpr) · GitHub Blog · 4 upvotes · 0 comments - [We halved our continuous integration pipeline](https://daily.dev/posts/we-halved-our-continuous-integration-pipeline-w7assusoy) · LeadDev · 0 upvotes · 1 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#cicd](https://daily.dev/tags/cicd), [#circleci](https://daily.dev/tags/circleci) [View this post on daily.dev](https://daily.dev/posts/work-in-progress-wip-is-the-new-software-delivery-bottleneck-7xif44njt) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Work in progress (WIP) is the new software delivery bottleneck","url":"https://daily.dev/posts/work-in-progress-wip-is-the-new-software-delivery-bottleneck-7xif44njt","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/work-in-progress-wip-is-the-new-software-delivery-bottleneck-7xif44njt"},"datePublished":"2026-10-30T22:58:50.989Z","dateModified":"2026-10-30T22:59:20.035Z","description":"Coding agents let teams open far more branches at once, but median throughput on main hasn't improved. CircleCI's 2026 State of Software Delivery Q3 Pulse...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/c3ced8106c695f97edbc257aaadaf655?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/c3ced8106c695f97edbc257aaadaf655?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"CircleCI","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"CircleCI","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/661fc536429943fdb5426b9f33fd8e3c","url":"https://daily.dev/sources/circle"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/work-in-progress-wip-is-the-new-software-delivery-bottleneck-7xif44njt","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,cicd,circleci","timeRequired":"PT8M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"CircleCI","item":"https://daily.dev/sources/circle"},{"@type":"ListItem","position":3,"name":"Work in progress (WIP) is the new software delivery bottleneck"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/work-in-progress-wip-is-the-new-software-delivery-bottleneck-7xif44njt#faq","mainEntity":[{"@type":"Question","name":"why isn't shipping speed increasing even though my team uses more coding agents and opens more branches","acceptedAnswer":{"@type":"Answer","text":"Work in progress has become the bottleneck, not coding capacity. CircleCI's Q3 2026 data shows active feature branches per contributor rose from 7 to 9 and active projects grew 16%, yet median main-branch throughput stayed flat at 1.7 runs per day, since validation and review capacity didn't grow at the same rate as parallel work. Teams debugging stalled delivery despite more AI coding agents can track these same bottleneck patterns via daily.dev."}},{"@type":"Question","name":"how does Little's Law explain why more parallel development with coding agents slows down delivery","acceptedAnswer":{"@type":"Answer","text":"Little's Law states average cycle time equals average work in progress divided by average throughput. If agents double the number of changes in progress from 10 to 20 while merges per day stay at 5, cycle time doubles from 2 to 4 days; only matching the completion rate to the new WIP level keeps cycle time stable. Engineers reasoning about agent-driven workflow capacity can dig deeper into these delivery models on daily.dev."}},{"@type":"Question","name":"what metrics distinguish high-throughput engineering teams from median teams when scaling coding agents","acceptedAnswer":{"@type":"Answer","text":"The top 5% of organizations by main-branch activity grew active feature branches by 55% while also increasing main-branch throughput by 31%, widening their lead over the median team's flat throughput to more than 10 times by Q3 2026. Key tracked metrics include WIP, cycle time, time to first review, and merge efficiency ratio (CI runs on branches relative to main). Teams benchmarking delivery velocity against peers can follow this kind of throughput data on daily.dev."}}]} ```