Blog
AI Articles & Insights
Writing on AI visibility, LLM search behaviour and how brands show up in generative answers.
Gemini 3.7 Flash: What a Sharper Coding Model Means for AI Answers
On 13 August 2026 Google made Gemini 3.7 Flash generally available — a sharper coding-and-agents workhorse. Here’s what it means for your AI visibility.
This Week in AI: Agents Go to Work Across the Enterprise
In the week to 15 August 2026, the leading AI platforms moved from proof to deployment, leading new pages on enterprise adoption, agents in the browser and a sharpening frontier-security posture.
This Week in AI: The Labs Turn Trust Into Enterprise Reach
In the week to 8 August 2026, OpenAI and Anthropic turned the trust story into commercial reach: named customers and partners, plugin and connector distribution and hands-on enablement led the new pages, while trust, safety and governance work continued as a steady thread across the leading AI platforms.
This Week in AI: Safety Takes the Front Page as Ecosystems Widen
In the week to 1 August 2026, OpenAI published a large library of AI threat-disruption case studies while OpenAI and Anthropic kept widening their partner and plugin ecosystems - trust and safety moved to the front page.
Sign in with ChatGPT: What the Identity Play Means for AI Answers
On 29 July 2026, OpenAI began rolling out Sign in with ChatGPT across partner apps. What ChatGPT becoming an identity layer means for how AI answers name your brand.
Claude Opus 5: What a Near-Frontier Model Means for AI Answers
On 24 July 2026, Anthropic launched Claude Opus 5 — near-frontier reasoning at half the price. Here’s what a stronger, cheaper Claude means for your brand.
This Week in AI: The Plugin Shelf Fills, Security Wires In
In the week to 25 July 2026, Anthropic filled the Claude plugin shelf while OpenAI widened its partner network and launched OpenAI Presence - with agent security wiring in underneath.
ChatGPT Health: What a New Health Surface Means for AI Answers
On 23 July 2026, OpenAI launched ChatGPT Health, a dedicated health experience in ChatGPT. Here’s what a new health answer surface means for your brand.
ChatGPT Ads: What Shorter Answers Mean for AI Visibility
On 24 July 2026, reconnAI tracked data shows the average ChatGPT answer is less than half the length it was in January — with the sharpest falls landing as ChatGPT ads roll out. What shrinking answers mean for whether AI names your brand.
Gemini 3.6 Flash: What Google’s Upgraded Model Means for AI Answers
On 21 July 2026 Google made Gemini 3.6 Flash generally available — a faster, cheaper, upgraded everyday model, live across the Gemini app and API worldwide. What it means for whether AI names your brand.
ChatGPT Triples Its Custom Instructions Limit to 5,000 Characters
On 15 July 2026 ChatGPT raised its custom instructions limit to 5,000 characters — more than triple the old cap. What deeper personalisation means for whether AI names your brand.
ChatGPT Desktop: What Chat and Work Mode Means for AI Answers
On 16 July 2026 ChatGPT’s desktop app split into Chat and Work with a ChatGPT and Codex switcher — what making agentic Work a default surface means for AI answers.
This Week in AI: OpenAI Signs Up the Consultancies
In the week to 18 July 2026, OpenAI published more than forty consulting and integrator partner pages while Anthropic stacked up frontier-customer proof and opened education as a new front. The new pages point squarely at the enterprise deployment channel.
ChatGPT on WhatsApp: What the EEA Return Means for AI Answers
On 13 July 2026 ChatGPT returned to WhatsApp across the EEA — no account needed. What a two-billion-user messaging surface changes for AI answers.
Perplexity Brain: What Agent Memory Means for AI Answers
On 13 July 2026 Perplexity shipped Brain, memory that learns from every task, plus one-click website publishing. What persistent agent memory changes for AI answers.
ChatGPT Sites: What AI-Built Pages Mean for AI Answers
On 9 July 2026 OpenAI opened ChatGPT Sites in public beta — publish a website straight from ChatGPT. What AI-built pages mean for AI answers.
ChatGPT Work: What OpenAI’s New Agent Means for AI Answers
On 9 July 2026 OpenAI launched ChatGPT Work — an agent that researches across your apps and builds finished reports. What it changes for AI answers.
ChatGPT Voice Gets GPT-Live-1: What It Means for AI Answers
On 8 July 2026 ChatGPT Voice switched to GPT-Live-1, a new model that searches the web mid-conversation. Here’s what it changes for how brands get named.
This Week in AI: The Labs Wire Into the Enterprise Tool Stack
In the week to 11 July 2026, OpenAI and Anthropic filled out plugin and connector directories, pushed Cowork to web and mobile and added enterprise proof.
ChatGPT’s New Fallback Model: What GPT-5.5 Instant Mini Changes
ChatGPT swapped in GPT-5.5 Instant Mini as its fallback model on 6 July 2026. Here’s what the model you reach at your limit means for AI answers.
This Week in AI: Labs Build Out Verticals as a Science Front Opens
In the week to 4 July 2026, Claude and OpenAI leaned into verticals, a wider integration marketplace and agentic workspaces - with a new science front.
Claude Sonnet 5 Launches: What the New Model Means for AI Answers
Anthropic launched Claude Sonnet 5 on 30 June 2026 — its most agentic Sonnet yet. Here’s what the new model changes for how AI answers name brands.
ChatGPT Retires GPT-4.5: What the Final Move to GPT-5.5 Means for AI Answers
On 26 June 2026, OpenAI retired GPT-4.5 from ChatGPT — the last model of the GPT-4 generation — and routed those conversations, and any custom GPTs pinned to it, onto GPT-5.5. Swapping the model behind an answer can change which brands and sources it names. Here’s what the retirement changes and why it’s a re-baseline moment for AI answers.
This Week in AI: Ecosystems Widen as OpenAI Opens a Security Front
In the week to 27 June 2026, OpenAI and Anthropic widened the tooling around their models — stacking up integrations, giving agents enterprise job titles across finance, sales and engineering, and, new this week, opening a serious front on AI security and trust.
ChatGPT’s GPT-5.5 Instant Levels Up Recommendations: What It Means for AI Answers
On 24 June 2026, OpenAI updated GPT-5.5 Instant — ChatGPT’s most-used model — to be sharper at decisions, shopping and “which should I pick” questions. When the model that answers most buying questions changes how it recommends, the brands it names change with it.
Perplexity Levels Up Deep Research: What It Means for AI Answers
On 19 June 2026, Perplexity rolled a more powerful Deep Research into its Computer assistant — with forking, inline actions and enterprise controls. When an answer engine researches deeper, the brands it surfaces are chosen differently.
This Week in AI: Labs Race to Wire Up, Skill Up and Prove It
In the week to 20 June 2026, the leading AI platforms shipped far more plumbing than headlines. Across Claude, OpenAI and Perplexity, the new pages cluster around three priorities — wiring assistants into existing tools, packaging agents and skills, and stacking up customer proof by industry. Here is the direction of travel.
ChatGPT Ads Manager Beta Opens in the UK: What It Means for Your AI Visibility
On 19 June 2026, OpenAI opened its ChatGPT Ads Manager Beta to advertisers in the United Kingdom — a self-serve desk at ads.openai.com to build, run and report on campaigns inside ChatGPT. Two weeks after ads started appearing to UK users, the tooling to buy them is now live. Here is what that shift means for brands competing to be seen in AI answers.
Claude Arrives in Microsoft 365 Copilot: What It Means for AI Answers
On 16 June 2026, Microsoft added Anthropic’s Claude as a selectable model inside Microsoft 365 Copilot Chat. It looks like one more option in a dropdown, but the model behind an answer shapes which brands and sources get named — and Copilot is where millions of employees ask their questions.
Google Clarifies Its llms.txt Guidance: What It Means for Your AI Visibility
On 15 June 2026 Google confirmed that llms.txt files aren’t used by Google Search and won’t help or hurt your rankings or visibility. Here’s what that clarification means for how brands get found in AI answers — and where to focus instead.
Claude Fable 5 and Mythos 5 Suspended: What It Means for AI Answers
On 12 June 2026 Anthropic suspended Claude Fable 5 and Mythos 5 just three days after launch, reverting Claude to Opus 4.8. Here is what the rollback means for your brand in Claude’s answers — and why now is the time to re-baseline.
Is Your Brand Ready for AI Search?
A strategic guide for marketing leaders: where GEO fits in your mix, how to tell if your brand is a fit for AI search, and when to start investing.
How LLMs Decide Which Brands to Cite
What the evidence shows about how ChatGPT, Gemini and Perplexity choose which brands to cite, and why strong SEO foundations still matter.
GPT-5.2 Is Retired: What ChatGPT's Move to GPT-5.5 Means for AI Answers
On 12 June 2026, OpenAI retired the GPT-5.2 family from ChatGPT and now routes those conversations onto GPT-5.5. It looks like housekeeping, but swapping the model behind an answer can change which sources and brands get named. Here is what the retirement changes, why the 90-day model lifecycle matters, and what to watch as answers re-settle on GPT-5.5.
OpenAI Files for IPO and Enters Advertising — While Anthropic Expands Its Enterprise Platform at Record Scale
This week OpenAI quietly filed a confidential S-1, launched ad-tool policies and acquired ONA, while Anthropic published more than 300 new pages across connectors, managed agents, international markets and Cowork. Gemini and Perplexity were silent.
ChatGPT's Simplified Model Picker: What the New Reasoning Tiers Mean for AI Visibility
On 10 June 2026, ChatGPT simplified its model picker — Instant, Medium, High and Extra High replace the old Thinking tiers, Thinking-Light is gone, and Instant can now auto-escalate to Medium when a question needs it. Here is what the new reasoning controls change about how everyday questions get answered.
ChatGPT Lockdown Mode: A Lowdown on the Lockdown
OpenAI shipped Lockdown Mode to every ChatGPT user on 4 June 2026 with little fanfare — and most coverage misread it. An early, accurate read on what it actually defends against, what it deliberately does not, and why it matters before everyone else catches up.
Claude Fable 5 Launch: Key Model Findings
Anthropic launched Claude Fable 5 on 9 June 2026 — a new Mythos-class model released for general public use alongside Claude Mythos 5. Here are the key findings from the launch: what the new generation is, how it compares to the Opus 4.x line, and why its broad availability is the real headline.
Building a GEO Strategy From the Ground Up
Learn how to build a rigorous Generative Engine Optimisation (GEO) strategy. Our practical framework helps brands track, measure, and improve their visibility across AI search platforms like ChatGPT and Google AI Overviews.
ChatGPT Ads Roll Out in the UK: What It Means for Your AI Visibility
OpenAI has begun showing ads inside ChatGPT to Free and Go plan users in the UK. Here is what the shift means for British brands - paid-versus-organic competition, the sectors most exposed, and how reconnAI tracks sponsored placements inside LLM answers.
GPT-5.5 Instant Response-Style Update: What Shorter, Canvas-Free Answers Mean for Your AI Visibility
On 28 May 2026 OpenAI refreshed GPT-5.5 Instant in ChatGPT and the API — shorter, more conversational answers, fewer bullet-heavy lists, and canvas retired for inline writing and code blocks. Here’s what the new default response style means for your brand’s visibility in AI answers, and what reconnAI customers should track now.
Google Preferred Sources in AI Overviews & AI Mode: What It Means for Your AI Visibility
On 27 May 2026 Google extended its Preferred Sources feature to AI Overviews and AI Mode, letting people pin the sites they want prioritised in AI answers. Here’s how that reshapes brand visibility in Google’s AI surfaces — and what reconnAI customers should track now.
Claude Opus 4.8 Launch: What It Means for Your AI Visibility
Anthropic released Claude Opus 4.8 on 28 May 2026 with stronger coding, agentic skills, reasoning and knowledge work over Opus 4.7. Here’s how the upgrade reshapes brand visibility inside Claude — and what reconnAI customers should re-baseline now.
Google Spam Policies Now Apply to AI Responses: What It Means for AI Visibility
On 15 May 2026 Google confirmed that its Search spam policies apply to AI-generated responses, including AI Overviews and AI Mode. Here is what the clarification means for AI visibility tracking, which tactics are now explicitly off the table, and how brands should respond.
GPT-5.5 Instant in ChatGPT: What the New Default Model Means for AI Visibility
OpenAI replaced GPT-5.3 Instant with GPT-5.5 Instant as ChatGPT's default model on 5 May 2026, bringing sharper accuracy, tighter answers and stronger image and STEM reasoning. Here is what the default model swap means for AI visibility tracking, and what brands should be re-baselining now.
ChatGPT Memory Sources: What the May 2026 Update Means for AI Visibility
ChatGPT's 5 May 2026 memory sources update lets users see which past chats, files and Gmail emails shaped each response. Here's what it changes for AI visibility tracking and which signals brands should start monitoring now.
Track LLM Text Fragments in GA4: A 15-Minute Setup Guide
LLMs link to your site using URL text fragments. GA4 misses them. A 15-minute Google Tag Manager setup that captures the actual text being quoted.
ChatGPT Fast Answers: What Bypassing Memory Means for AI Visibility
On 22 April 2026, OpenAI rolled out Fast Answers in ChatGPT - a shortcut that skips memory and past chats to deliver quicker responses to factual queries. Here is what the change means for brand visibility inside ChatGPT, and how reconnAI is tracking the new answer pathway.
ChatGPT Ads Rollout in Australia, New Zealand and Canada: What It Means for AI Visibility
OpenAI has begun showing ads inside ChatGPT to Free and Go plan users across Australia, New Zealand and Canada. Here is what the shift means for brand visibility, competitor positioning, and how reconnAI tracks paid placements inside LLM answers.
Claude Opus 4.7 Launch: What It Means for Your AI Visibility
Anthropic released Claude Opus 4.7 on 16 April 2026 with sharper software engineering, longer-horizon reasoning, and higher-resolution image understanding. Here's how the upgrade reshapes brand visibility inside Claude.
When the World Shifts: How Geopolitical Events Quietly Reshape Your Brand’s AI Visibility
.article-hero-img{display:none !important;} Wars, sanctions, and global crises don't just dominate headlines - they subtly alter how AI models talk about your …
The Accuracy Imperative: Why Factual Precision Is Now the Defining Challenge for AI - and the Brands That Depend on It
.article-hero-img{display:none !important;} The race to build smarter, faster, more capable AI has dominated headlines for years. But as large language models …