Product Changelog
October 19, 2026
Insights Prompt Drilldowns and Combined Campaign Criteria
weekly-update-2026-10-19
Explore the prompts behind a query fanout and investigate citation declines from your connected assistant. Combine Campaign criteria to guide opportunity discovery, and prepare structured option lists for your team’s work in Grids.
- Trace query fanouts back to their originating prompts in Insights. Open a query’s detail panel to review the full matching prompt list, including topics, observation counts, platforms, and last observed dates, then narrow the results by date, platform, or topic.
- Investigate citation declines through your connected assistant with lifecycle metrics in AirOps MCP. Filter and sort cited URLs by their decline from peak, then request a URL’s weekly citation history to examine changes, including weeks with no citations, within your selected period.
- Combine Campaign criteria to reflect your team’s priorities during setup or in Campaign settings. Select multiple available criteria templates, adjust their values, and add freeform guidance alongside them to describe which opportunities AirOps should surface for the chosen action type.
- Convert existing Grid data into Single Select options from populated Text or JSON columns. Preview detected options and affected cells, then rename or merge options before converting; editing supports up to 200 detected options, and the preview identifies cells that would remain unselected.
- Prepare option lists from Workflow results with the new Format Single/Multi Select Step. Choose the fields for display labels and stored values, then output JSON or CSV, such as readable author names paired with CMS IDs, for use when configuring Grid options.
October 12, 2026
Query Fanouts and Google AI Mode Ads
weekly-update-2026-10-12
Explore the searches behind your tracked prompts and the ads appearing in Google AI Mode answers. New MCP tools bring content feedback and output review into connected assistants, while Claude Haiku 5.5 adds another model choice for Workflows.
- Find content gaps in the searches behind your prompts with Query Fanouts in Insights. Review the additional queries ChatGPT uses while researching your tracked questions, then compare recurring searches to identify topics to track or details your content should address.
- Research Google AI Mode ads in Ad Library alongside observed ChatGPT ads. Browse advertisers and copy, then open an ad or associated answer to inspect its context; coverage reflects the ads AirOps captures for your tracked prompts.
- Review content feedback from your connected assistant through AirOps MCP. Read Artifact comments with highlighted text and surrounding sentences, then inspect the Playbook instructions to understand the feedback and decide where a change belongs before editing.
- Test Playbooks and inspect their outputs through MCP from Claude or Codex. Run a draft or chosen published version and read the resulting output to check whether changes addressed the feedback; you can also find published Workflows, run them, and check results.
- Choose Claude Haiku 5.5 for lightweight Workflow tasks in Prompt LLM Steps, including classification and extraction. Enable adaptive thinking with Low, Medium, or High effort when the task needs more reasoning, and use vision or Web Research for supporting context.
October 5, 2026
Citation Lifecycle and Agent Analytics in Connected Assistants
weekly-update-2026-10-05
See how individual URLs gain and lose citations, and bring AI crawl activity into your team's analysis. Citation lifecycle reporting and Agent Analytics access help you investigate content opportunities, while GPT-6.1 Sol adds a new model choice for Workflows.
- Evaluate citation momentum before choosing where to focus with lifecycle metrics in Insights. Review when a URL first appeared, its peak week, rise, decline from peak, and confirmed half-life to investigate whether a potential content opportunity is gaining or losing citations.
- Inspect and export a URL's citation history by opening its weekly chart and reviewing the associated prompts. Carry lifecycle metrics into a CSV or Grid to compare opportunities, with the selected dates and filters defining the scope of your analysis.
- Analyze AI crawl activity through your connected assistant with Agent Analytics data in AirOps MCP. Once Agent Analytics is connected, compare AI Agent Visits, training crawls, and web search crawls with citation data to investigate pages that agents reach but rarely cite.
- Build custom reports from daily Agent Analytics data through the API or MCP. Filter counts by platform, crawl type, or a tracked page, then use the returned daily records to examine crawl patterns and AI Agent Visits over your chosen period.
- Use GPT-6.1 Sol in Prompt LLM Steps for new or existing Workflows. New Steps select it by default with Medium reasoning; choose Low or High to adjust the reasoning effort for your task, with support for web research and structured JSON output.
September 28, 2026
Ads Insights plus New OpenAI and Anthropic Models
weekly-update-2026-09-28
Compare how AI platforms cite YouTube content and explore observed ChatGPT ads in a visual gallery. New model choices also give your team more options for complex content work and faster execution.
- Compare YouTube citations across AI platforms in the Social tab. Switch between citation share and the percentage of tracked answers citing YouTube to examine different measures of its presence, with charts that follow your Insights filters.
- Explore the video characteristics behind YouTube citations through breakdowns by format, duration, age, channel size, or sponsorship signals. Review the underlying counts in chart tooltips to understand which groups contribute to the pattern before choosing content to investigate.
- Browse observed ChatGPT ads in the Ad Library gallery to compare creative and copy side by side. Open an ad to inspect its destination, appearance frequency, and the tracked prompts and topics where it appeared, then use those observations to inform your team's messaging.
- Choose GPT-6 Sol or GPT-6 Luna for content work in Playbooks and Workflows. Select Sol for complex tasks or try Luna when speed is the priority, then compare results on the work your team needs to complete.
- Create with Claude Opus 5.5 in Playbooks and Workflows when your team needs another model option for complex work. Choose it in the editor and compare its outputs on your team's tasks before settling on a model for the process.
September 21, 2026
Campaigns Updates, On-Demand Page Tracking, and Connected Content Reviews
weekly-update-2026-09-21
Reuse Campaign setups and get model-powered help while building Workflows. Content review notifications and on-demand page captures also keep your team connected to the work awaiting a decision and the changes that have reached your site.
- Duplicate a Campaign to reuse its setup and criteria for another effort. The copy starts as a draft with its Measurement structure intact and no generated opportunities or Action rows, so your team can adapt the scope before publishing and generating fresh work.
- Build and improve Workflows with Claude Sonnet 5 in Copilot, which uses the model by default. Ask Copilot to explain how a Workflow operates, suggest Steps and connections, or help diagnose a problem while you decide which changes to make.
- Send content review notifications to private Slack channels where your reviewers already collaborate. Invite the AirOps bot to the channel and select it for notifications, keeping review requests within the appropriate team while work waits for a human decision.
- Capture page changes when your team needs to review them with Check for changes on an enrolled page. Request a fresh capture after an edit, then inspect its change history to review the changes that have reached your site.
- Give long reporting details more room in Insights with resizable table columns. Adjust column widths when reviewing page titles, prompts, or citation details so the information your team needs is easier to read during analysis.
September 14, 2026
Social & Community Insights, Prompt Changelog, and Editable Campaign Criteria
weekly-update-2026-09-14
Explore the YouTube videos and Reddit discussions that appear in AI answers, and understand changes to the prompts behind your reporting. New creation models and editable Campaign criteria give your team more ways to turn those insights into work.
- Explore Social and Community insights to see which YouTube creators, videos, Reddit threads, and subreddits appear in tracked AI answers. Compare brand presence and review cited sources to identify content topics, creator partnerships, or community discussions your team should investigate.
- Understand reporting changes with Prompt Change Log and annotations on Visibility charts. Review which prompts or assignments changed, when they changed, and who edited them; customizable Insights table columns help you keep the metrics and context you use most in view.
- Create with GPT-6 Astra and GPT Image 2.5 in your content processes. Astra supports complex reasoning and web research in Workflows, while Sunburst and Flare provide image generation and editing with quality and dimension controls for the assets your team needs.
- Research ChatGPT ads and advertisers through MCP from your connected assistant. Inspect creative details and appearance patterns for tracked prompts without connecting an Ads account, using the ads AirOps observes to investigate competitor messaging and relevant advertising activity.
- Refine Campaign opportunity criteria after publishing as your team's priorities change. Update what to prioritize, exclude, or rank higher, then use the saved criteria for future opportunity generation while keeping existing opportunities and accepted work intact.
September 7, 2026
Prompt Share of Voice, Regional Tracking, and Content Review
weekly-update-2026-09-07
Choose from more models and pass the right Campaign context into Workflows and Playbooks. Grid controls, prompt reporting, and Session review details make it easier to configure work and inspect its results.
- Choose additional Claude and Gemini models for Workflows, including Claude Fable 5.1 and Gemini 3.7 and 3.8 Flash. The API Step also supports PUT and PATCH, allowing Workflows to update external records using the method the service requires.
- Connect Campaign context and Playbook outputs to downstream work by mapping the target page, brief, or supporting context into inputs. MCP also supports mapping Playbook Artifacts into Grid columns, keeping generated outputs available for the next stage of review or execution.
- Review prompt Share of Voice and refine tracking assignments to understand your brand's share of mentions for individual questions. Topic and platform filters help you narrow the tracked set before changing assignments, keeping your analysis focused on the markets and questions that matter.
- Follow Content Review progress within Playbook Sessions, with reviewer activity grouped under the check that produced it. Feedback suggestions also check the latest Playbook draft before applying an edit, helping protect changes your team has already made.
- Expand prompt tracking to additional regions, including Saudi Arabia, Taiwan, Kenya, Nigeria, and South Africa. Use the added region options to build a view of the markets your team serves and compare results across locations, with collection coverage varying by answer engine.
August 31, 2026
Prompt Intent Filters and Interactive AirOps Analysis in Codex
weekly-update-2026-08-31
Compare AI search performance by buyer intent and bring AirOps Insights into Codex tasks. These updates help your team investigate where its brand appears, compare results, and review changes to brand guidance in the same conversation.
- Compare AI search performance by buyer intent with Prompt Intent filters in supported Insights views. Separate awareness and consideration questions from evaluation and purchase questions to see how results vary across stages of the buying process.
- Investigate brand presence and citations within an intent stage by reviewing the prompts behind that segment. Use those comparisons to decide which topics or questions need closer attention and keep your analysis focused on a specific part of the buyer journey.
- Explore AirOps reporting inside Codex tasks through interactive charts and data tables. Ask your connected assistant to compare results by competitor, topic, or platform, keeping the data and the discussion together as you investigate.
- Refine an analysis within the same Codex conversation by changing the date range or metric as questions emerge. Follow up on a comparison to examine a narrower period or another measure without starting a separate reporting discussion.
- Review Brand Kit edits directly in Codex with interfaces that show proposed changes inside your task. Accept individual suggestions before applying them, so your team can update the guidance it wants while retaining control over the final edits.