Following last week's investment summit in Manhattan – hosted by this news service and featuring a raft of high-profile figures – we carry the first in a round of reports about the conversations.

Here is the first of a series of reports stemming from the conversations held at this publication’s FWR investment summit in New York City. (For a report giving an overview of the event, from our US correspondent Charles Paikert, see here.)

The following article is from Anna Garcia, who is the founder and managing partner of Altari Ventures.
 

Anna Garcia (left) and Tamara Zubatiy Nelson (right).

At the recent Family Wealth Report Investment Summit, I sat down with Tamara Zubatiy Nelson, Forbes 30 Under 30 CEO and founder of Barometer, an Altari Ventures portfolio company building AI infrastructure to safeguard brand reputation. Our conversation explored the central challenge facing enterprises today: how to harness AI's transformative potential without sacrificing the brand trust that underpins their entire business.

As an enterprise fintech investor, I observe a clear paradox: the AI technologies most capable of revolutionizing financial services simultaneously introduce some of the biggest risks.

Financial institutions spent $35 billion on AI in 2023, with projections reaching $97 billion by 2027 – the fastest growth of any major industry. AI is being deployed across everything from workflow optimization and analysis to content generation and monitoring systems designed to reach customers through every conceivable channel. 

McKinsey estimates generative AI could add $200 to $340 billion in annual value to global banking. Yet unlocking this value hinges on resolving a critical bottleneck: establishing trust in AI output at enterprise scale.

The gap between ambition and execution is staggering. Today, 75 per cent of financial firms deploy AI, yet only 12 per cent have implemented risk management frameworks. This chasm between adoption and governance represents an enormous emerging investment opportunity. The AI explainability and observability market is projected to exceed $21 billion by 2030, encompassing the infrastructure that makes AI decisions transparent, auditable, and defensible.

Barometer is building precisely this safeguards infrastructure – enabling brands to monitor and protect their reputational capital as they engage customers across expanding content channels. While Barometer serves major enterprises spanning multiple industries, financial institutions comprise their most substantial client base, validating both the pressing need and the solution's effectiveness.

From human-centered AI research to market need
Zubatiy Nelson's path to founding Barometer began during her PhD program at Georgia Tech, where she focused on human-centered AI – engineering systems to work for us rather than requiring us to adapt ourselves to work for AI. Around 2020, she identified a seismic shift in information consumption: sources were multiplying rapidly, moving from traditional one-to-many communication to individualized echo chambers – podcasts, YouTube channels, niche social platforms. Understanding how and where audiences consumed information became exponentially harder.

Simultaneously, advertising faced a reckoning with this transformed landscape. How could brands reach audiences effectively and safely amid such fragmentation? With her expertise in natural language processing, Zubatiy Nelson saw an opening for a company that could ingest this deluge of media, extracting coherent signals, and guiding brand activity in alignment with established values and risk parameters.

Today, Barometer functions as the premier brand suitability and contextual targeting platform for major purchasers of premium media. The system analyzes content before advertisements deploy, ensuring regulated industries like financial services avoid topics deemed inappropriate or illegal – unauthorized investment advice, gambling promotions, or worse. Integrated directly into the programmatic advertising ecosystem, Barometer operates as an oracle within the real-time bidding infrastructure, determining instantaneously where content should or shouldn't be placed.

Reputational risk at AI speed
The velocity of reputational risk has transformed dramatically with AI. One bad association or false piece of information can snowball into a full-blown brand crisis within minutes, with consequences that prove devastatingly punitive. Organizations cannot address this challenge through human vigilance alone – detection, monitoring, and response at human speed prove categorically insufficient against threats moving at AI speed.

Zubatiy Nelson framed the challenge as a dual evolution: both the problem and the solution space are changing in parallel. On the problem side, content channels continue to proliferate alongside explosive growth in volume itself – across user-generated material, premium media, podcasts. 

On the solution side, there's a critical distinction between traditional machine learning and what we're now calling AI. Machine learning operates deterministically: input X always yields the same result. Query ChatGPT with the same question twice, however, and you won't get identical answers. For a financial brand's reputational risk profile, that deviation is unacceptable.

An MIT study revealed that only 5 per cent of AI deployments in enterprise have progressed beyond pilot stage to scaled production delivering measurable value. This statistic illuminates the AI governance opportunity: establishing rigorous validation processes – both human oversight and technical infrastructure –enables organizations to achieve the same confidence in AI systems that they currently maintain in proven machine learning solutions.

AI governance as innovation enabler
For executives who hear "AI governance" and reflexively think "compliance checkbox" or "innovation impediment," Zubatiy Nelson articulated a different framing: AI governance represents a tremendous opportunity to unlock operational value and enable AI to deliver its promised business impact.

Consider Barometer's current approach: their machine learning system generates signals identifying which sites, videos, or social content align with client standards. Today, a human still reviews every single one before implementing changes. The governed AI future? Systems that automatically adjust, recognize anomalies, self-correct, report deviations – all operating with sufficient systemic trust that human intervention becomes unnecessary for routine determinations.

The governance infrastructure won't constitute mere oversight layered atop existing technology; it demands holistic reimagination of how these systems are architected, validated, and deployed to establish appropriate degrees of institutional confidence.

What's on the horizon
Looking ahead, Zubatiy Nelson described something approaching a renaissance in content creation and human potential. AI's promise of radical personalization is fueling an explosion of diverse perspectives and creative output. More profoundly, this evolution is unlocking human capabilities that previously lacked sufficient opportunity for expression.

The transformation unfolds across two dimensions. The immediate phase manifests as augmented productivity – employees wielding AI tools to amplify their effectiveness within existing roles. The more consequential shift, however, will emerge when enterprises redesign their core offerings around AI capabilities rather than merely enhance existing products. That inflection point – when AI moves from productivity enhancer to product foundation – represents where genuine value creation occurs.

Parting thoughts
AI can function as either hero or villain for brand reputation, particularly within regulated industries where trust constitutes a foundational product itself. It accelerates processes and enhances capabilities, but a single misstep can cascade into existential crisis within minutes. At the ambitious scale stakeholders envision for AI adoption, humans simply cannot maintain oversight without technological governance infrastructure.

Organizations that recognize AI governance as foundational architecture rather than ancillary consideration position themselves to capture disproportionate value. For investors, this represents an incredibly compelling opportunity: with billions of dollars to be invested to resolve the AI governance bottleneck, backing private early-stage companies building these solutions has the potential to create generational wealth.