OpenAI Presence Is Here. Your Competitor's Support Team Just Changed.

What Happened

On 22 July 2026, OpenAI launched Presence: an enterprise-grade AI agent platform for deploying voice and chat agents across customer-facing and internal workflows. It is available now, but only through a limited general availability programme, led by OpenAI Forward Deployed Engineers and select systems integrators. This is not a self-serve tool.

The headline number from OpenAI's own deployment: Presence now resolves 75% of inbound support issues without human assistance, and its Codex-powered improvement loop cut human handoffs by 15 percentage points in just 10 days. Those are not benchmarks from a controlled trial. That's what it did in production, on OpenAI's own phone support line.

How It Works (and Why That Detail Matters)

Presence isn't a chatbot builder. Each deployment starts with a single, defined job — resolving billing disputes, handling insurance claims, fielding IT service requests. The agent gets only the knowledge and system access it needs for that one job. Nothing more.

Companies set the policies: what the agent can do autonomously, when it needs approval, and when a human must take over. After launch, production sessions and escalations surface gaps, and Codex (OpenAI's coding and reasoning model, used here via a Presence plugin) proposes updates. Teams test those updates against the live version before approving any rollout. The loop is: deploy, observe, propose, test, approve.

That architecture — scoped access, policy-governed behaviour, continuous Codex-driven iteration — is what separates Presence from a generic LLM wrapper. It's designed for the moment after you've proven the concept and need it to work reliably at scale.

The Competitive Timeline: Who's Already Building on This

BBVA is exploring voice support for banking in Mexico. SoftBank is testing natural Japanese-language customer conversations. IAG is deploying Presence to support insurance customers during high-demand events like severe weather. These are not pilots announced two years out. These are live or near-live deployments from the launch announcement itself.

Table 1: OpenAI Presence — Early Enterprise Deployments (announced 22 July 2026)

Organisation	Sector	Use Case	Status
BBVA Mexico	Financial Services	AI-powered voice support for everyday banking	Exploring / Design Partner
SoftBank	Telecoms	Natural Japanese-language customer conversations	Testing
IAG (Retail Insurance, Australia)	Insurance	Claims support during high-demand events	Exploring
OpenAI (internal)	Technology	English-language phone support (1-888-GPT-0090)	Live — 75% resolution rate

Why This Is a Search and Discoverability Story, Not Just a CX Story

Here's the angle that most growth teams will miss: Presence changes where buying decisions get resolved, not just supported. When a prospect can speak to a voice agent that understands their billing question, verifies their account, applies company policy, and takes an approved action — all before they ever speak to a human — the traditional "content nurtures, sales closes" funnel compresses significantly.

For B2B SaaS teams investing in SEO and AI search visibility, that creates two connected pressure points. First, if your competitors deploy Presence and yours don't, their time-to-resolution drops and your churn risk increases. Second — and this is the GEO angle — AI agents like Presence are trained on, and will increasingly surface, vendor content that exists in structured, trustworthy, retrievable formats. If your product documentation, help content, and policy pages aren't optimised for AI retrieval, Presence-style agents at your customers' organisations may simply not find you when they go looking for solutions.

We've been advising clients on this shift for the past 18 months: the firms winning in AI search aren't just the ones with the best blog content. They're the ones whose content is structured so an AI agent can confidently act on it. This launch accelerates that requirement.

What B2B SaaS Growth Teams Should Do Now

First, treat Presence as a signal about where enterprise buyers are heading, not just a vendor announcement. The fact that BBVA, SoftBank, and IAG are all named at launch tells you something about the pace of enterprise adoption. If your ICP includes large-enterprise or regulated-industry accounts, your buyers are almost certainly evaluating this category.

Second, audit your content for AI retrievability. Help docs, pricing pages, integration documentation, and policy pages should be structured so that an AI agent — whether it's Presence, a ChatGPT-based tool, or a Gemini-powered assistant — can parse them accurately. That's not an SEO-as-usual task. It sits at the intersection of generative engine optimisation (GEO) and content architecture, which is where we're spending most of our advisory time right now.

Third, consider what Presence-style agents mean for your own support and sales motion. If 75% of support queries can be resolved without human involvement, the humans freed up should be redeployed into higher-value pipeline conversations. That's a capacity and positioning decision, not just an ops one.

Presence is available to eligible enterprise customers today through OpenAI's limited GA programme. If you want to explore it, you'll need to contact your OpenAI account team directly. For the broader strategic question of how this reshapes your discoverability and pipeline architecture, we've been mapping that terrain and we're happy to think it through with you.

Cover photo by Steve A Johnson on Unsplash