UiPath and integration-software vendors are on the wrong side of GPT-6 Astra's computer-use ability. OpenAI's Greg Brockman argues models that operate software like a person will bypass the APIs, plugins and RPA glue companies pay for. Astra scored 72.6% on OSWorld. If that holds in production, integration budgets shift from middleware to agent sellers. UiPath won't go straight to zero, being embedded in controls, audit logs and multi-year contracts; eight quarters of renewals decide it. x.com  … Show more

Microsoft is a bigger near-term competitive threat to UiPath than Claude, a point that gets overlooked. Claude is the best standalone model, but large enterprises often block it over security, compliance and data governance, while Microsoft's Copilot is already embedded in daily tools and getting good enough. Microsoft's agentic tools are not yet on par with UiPath and the two are partners for now, and orchestration may leave room for UiPath if management executes. x.com  x.com 

UiPath will be among the AI winners because AI agents cannot replace everything. Companies still need deterministic automation that is reliable, controllable and free of constant token consumption, and UiPath connects that world with agentic AI. After Q2 FY27 the case is more convincing than ever. x.com 

UiPath may carry the highest business-model risk of any SaaS company because LLM agents with computer use compete directly with its robotic process automation. Rivals include Anthropic and OpenAI themselves. Its base is sticky for now, with 97% gross retention, regulated large-enterprise customers and long contracts, giving it time to sell AI tools. Still, it would be very surprising if gross retention stays in the high 90s, and the 2x EV/ARR multiple reflects the threat. x.com 

Holding ServiceNow into earnings looks unwise after OpenAI released an agentic automation tool meant to replace UiPath and ServiceNow, sending UiPath down 13%. x.com 

UiPath, Zeta Global and ServiceNow are discounted because the market treats AI as a pure threat rather than a product accelerator, as it once did with Google, now up 120% this year. They should rerate once investors see that AI makes their products better, faster and more valuable, and that AI still needs real enterprise workflows, data, compliance, automation and execution layers to be useful. x.com 

Enterprises' need for source control, CI, security and governance at agent scale probably makes Anthropic's Scion more of an opportunity than a threat for GitLab. Scion shows what starts to break when agents generate code at this scale, a problem GitLab is already working to solve; enterprises need not build their own Scion or replace existing workflows. Scion itself is not inherently bullish for GitLab, and OpenAI, Meta or Anthropic selling that full stack broadly would pose a competitive risk. … Show more

Block may be hurt rather than helped if small businesses get Muse-like AI agents, because such agents could destroy its software margins and value proposition, which sit on top of commodity hardware and payments. x.com 

UiPath's real story is distribution through Microsoft, not bots. Per @JasonL_Capital, it partnered with Microsoft to integrate security automation into Defender, Sentinel and Security Copilot. Sitting inside the Microsoft stack lets UiPath inherit Microsoft's enterprise footprint overnight, a sticky position, because enterprise software winners become part of the workflow rather than another tool on top of it. x.com 

AI is not killing software; it is killing the per-seat software model, which is why Salesforce, HubSpot, ServiceNow, Snowflake and Datadog are getting crushed. Companies will not keep paying for 100 Salesforce users, 50 HubSpot users and 30 ServiceNow users when five AI agents can do the same work around the clock. x.com