When the CEO Sent His Clone to the Earnings Call: AI Twins Go Corporate, and What That Means for Every Brand

The earnings call that changed the conversation

On April 24, 2026, analysts dialing into Customers Bank's first-quarter earnings call had no idea anything unusual was happening. The prepared remarks came through clear, calm, and precise. About 30 minutes in, CEO Sam Sidhu dropped the reveal: the voice they'd been listening to was not his. An AI clone, trained on his speech patterns, cadence, and language, had delivered the entire prepared section of the call.

"The prepared remarks you heard on my behalf today were delivered by my AI clone, not read by me," Sidhu told analysts, who by multiple accounts were genuinely surprised.

The move was deliberate theater with a strategic point behind it: Customers Bank, a $25.9 billion asset lender, is going all-in on AI-enabled operations. The bank has since announced a multi-year partnership with OpenAI and says it has already saved 28,000 hours of work, the equivalent of about 15 full-time employees, using AI across its workflows. Sidhu told CNBC he expects the project to improve the firm's efficiency ratio from roughly 49 percent to the low 40s.

The internal AI story is interesting enough. More interesting is what his choice to put a clone on the most scrutinized investor call of the quarter says about where executive communication is heading, and every brand's content strategy with it.

Sidhu isn't alone

The earnings call was dramatic, but the trend behind it has been building for months. Sidhu joins a list of high-profile executives investing in their own digital replicas.

Mark Zuckerberg is building an AI version of himself trained on his mannerisms, tone, public statements, and views on company strategy. The rationale is internal: Meta's 79,000 employees can't all get meaningful time with their CEO, so a digital twin answers questions and delivers messages in his place. Sebastian Siemiatkowski of Klarna and Eric Yuan of Zoom have announced digital twin projects of their own, each citing a version of the same logic: there is more demand for their voice than there are hours in their day.

What separates this moment from the earlier wave of "virtual CEO" experiments is execution quality. The clones being built in 2026 are not robotic text-to-speech avatars pasted over a stock headshot. They are full video and audio replicas, trained on the actual person's likeness, voice, speaking rhythm, and expertise, and they hold up under scrutiny from analysts, employees, and investors.

Why this matters beyond the C-suite

It's tempting to file the CEO digital twin story under things for Fortune 500 comms departments to worry about. That misses the real signal.

Large companies have always been able to produce executive content in volume: ghostwritten op-eds, video series, podcast appearances, speaker bureau placements, polished earnings decks. What they lacked was a way to personalize and scale that content across every touchpoint without a small army of producers. AI twins fix that.

Personal brands and independent creators have the opposite problem. Audiences genuinely want their voice, but there's no team, no production budget, and the same 24 hours everyone else gets. An AI twin doesn't replace that voice; it amplifies it. A founder with an audience of 50,000 can show up in their customers' feeds with consistent, high-quality video every day while doing the work that actually builds the business.

That's what Sidhu demonstrated on an earnings call, and what creators and brand builders are testing right now: the AI twin as a force multiplier for anyone whose voice has equity but whose time does not scale.

The new economics of presence

The economic logic gets clearer once you look at the numbers.

A traditional branded video presence, say a weekly founder-to-camera post, monthly long-form content, and ad creative refreshed every quarter, takes a serious ongoing investment in production, editing, and distribution. For a solo creator or a small brand, the bottleneck is not ideas. It's execution. You can't shoot five videos a week and run a company at the same time.

AI twin platforms in 2026, including HeyGen, D-ID, Synthesia, and several newer challengers, offer custom-trained replicas starting at roughly $300 to $1,200 per month at the professional tier. That includes a training session to capture your likeness and voice, plus generation capacity that turns a script into video in minutes rather than a production day. The twin works around the clock, needs no lighting setup, and never drifts off-brand.

The right comparison isn't a twin versus doing nothing. It's a twin versus a video production retainer, which for comparable output runs $3,000 to $15,000 a month at a boutique agency. The math is not subtle.

On the advertising side, the same economics apply to UGC and ad creative. At roughly $2 to $11 per AI-generated video, against $150 to $500 per video from a human creator, a small team can afford the kind of structured creative testing (dozens of hook variations, avatar types, scripts, and formats) that until recently only brands with large performance marketing budgets could run.

The authenticity question

There's an obvious objection, and it deserves a real answer. If an AI twin is doing the talking, is the presence authentic at all? The question got sharper after Sidhu's call, which some commentators framed as deceptive. Sidhu's answer was disclosure: he said, deliberately and publicly, exactly what he had done and why. The AI delivered the scripted, rehearsed portion of the call, the part that would have been identical whether he read it live or not. The judgment calls, the Q&A, the strategic decisions stayed his.

That distinction matters for anyone weighing an AI twin. The value of a personal brand was never the physical act of reading words to a camera. It's the ideas, the expertise, and the trust a person has built over time, and a twin delivers those at scale in that person's voice and image. It can only extend authority someone already has; it can't manufacture authority that isn't there.

Disclosure is non-negotiable. Meta, TikTok, YouTube, and LinkedIn all require labels on AI-generated content now, and that's the right call, ethically and strategically. Recent platform studies suggest transparent disclosure doesn't materially hurt engagement when the content is genuinely valuable. Sometimes the novelty even works in a brand's favor: audiences are curious about the technology, and creators who are open about it start conversations their competitors aren't having.

What the smartest brands are doing right now

The companies and creators getting the most out of AI twins in 2026 treat them as infrastructure, not as a cost-cutting shortcut.

The pattern that keeps surfacing in case studies and platform reports is a hybrid model. The human shows up for content that needs real-time judgment, emotional nuance, or a personal moment nobody can reproduce. The AI twin handles the volume layer: consistent weekly content, ad variations, multilingual versions, localized cuts for different markets, all the material that would otherwise require a team or simply not get made.

For a creator or founder who has built an audience but hit a ceiling on how much content they can personally produce, this isn't a theoretical future. It works today.

The Customers Bank call was surprising partly because of how polished it was, and partly because of how matter-of-fact Sidhu was about it afterward. No apology, no defensiveness. He framed it as a demonstration of a capability his bank already uses every day at scale. AI twin as standard operating procedure rather than exotic experiment: that framing is the part worth watching.

The bigger picture

Two weeks ago, the CEO of a $25.9 billion bank opened his earnings call with an AI clone. This week, Mark Zuckerberg is training a digital version of himself to stay connected to 79,000 employees. And the same technology behind those enterprise deployments is already available to independent creators, DTC brands, and small businesses whose voice needs to show up in more places than their schedule allows.

Read the executive AI twin story less as a warning about authenticity eroding and more as a proof of concept for what happens when physical presence stops being the constraint. Sidhu didn't use his clone because he had nothing to say. He used it because saying the right things consistently, at scale, is the work, and AI twins are now good enough to do part of that work well.

For brands and creators who have spent years wishing they could show up more consistently, reach more markets, and still have time to build their products, that constraint just got a lot smaller.