BOSTON, Sept. 14, 2026 (GLOBE NEWSWIRE) -- Anthropic CEO Dario Amodei’s call to slow the development of frontier AI models has received support from OpenAI CEO Sam Altman and Elon Musk. Nishant Gupta, CEO and Founder of OneMetric, believes the discussion must also address the models enterprises already depend on.
“Most AI users have encountered this: a model that previously worked reliably suddenly responds differently, even though the surrounding application remains unchanged,” says Nishant.
“The model may have been updated, rerouted, restricted, or optimized for cost. From the user’s perspective, it is difficult to determine what actually changed.”
This lack of visibility is especially concerning when a model is integrated with customer communications, CRM data, proprietary knowledge, or business decisions. Enterprises remain accountable for their applications, even with limited control over underlying changes.
“Adoption should not move faster than organisational readiness” is another principle Nishant has frequently emphasized in discussions with enterprise leaders.
He sees these two points as connected. Enterprises must establish the foundations needed for responsible AI use, while model providers must ensure the stability and transparency necessary to maintain that foundation.
Pacing Must Include Existing Models
In his essay, We Must Pace the Frontier, Amodei argues that AI capabilities should not outpace our ability to understand, evaluate, and control them. His focus is primarily on the next generation of advanced models.
Nishant believes this standard should also apply to models already in enterprise use. Pacing should involve greater stability, version transparency, clear communication of significant changes, and accountability for their impact on existing applications.
An enterprise may design an AI system based on a model’s observed behavior. If the model changes, the organization may assume new risks without altering its own systems.
Enterprise AI Needs a Business Foundation
Many enterprises already use approved LLMs, private deployments, or models integrated with internal systems. Additional model capability is not always what is lacking.
The more important question is whether the AI understands the organization’s operations.
Enterprise data alone is rarely sufficient. AI also requires business context, institutional knowledge, defined capabilities, and operating rules. It must recognize what has changed, why it matters, which expertise applies, what action is appropriate, and when human involvement is needed.
This foundation should not be rebuilt for each new agent. Once an organization’s data, context, expertise, skills, and governance are structured, the same foundation can support multiple AI solutions and future use cases.
This approach changes the value enterprises derive from AI.
Without a foundation, an agent may generate emails, enrich records, summarize conversations, or assign scores. While these are completed tasks, they do not necessarily represent business progress.
With the right foundation, AI can help identify the right account sooner, improve prioritization, select the next best action, prevent unsuitable activities, or create a more relevant customer experience. The output shifts from task completion to tangible business outcomes.
It also becomes easier to govern. Teams can define system access, decision-making authority, required confidence levels, when actions should be suppressed, and where human judgment is necessary.
This is where responsibility must be shared. Model providers should communicate changes that may affect existing applications. Enterprises must define the context and rules for AI operation. Application builders must translate these requirements into permissions, evaluations, escalation paths, and human checkpoints.
Pacing the next frontier is important. However, enterprise AI will only become dependable when existing models operate on foundations built for accountability and measurable outcomes.
About OneMetric
OneMetric is a HubSpot Elite Partner helping businesses solve complex GTM, RevOps, CRM, engineering, and digital transformation challenges. It has supported more than 750 customers across SaaS, fintech, healthcare, real estate, and other industries.
Its team includes 150+ HubSpot-certified experts and 15+ dedicated GenAI engineers. OneMetric is SOC 2 Type II and ISO 27001 certified.
Media Contact
OneMetric
Saurabh Kumar
https://www.onemetric.io/
saurabh.b@growtomation.in
Disclaimer: This content is provided by OneMetric. The publisher and GlobeNewswire do not independently verify, endorse, or assume liability for its accuracy, completeness, or claims.
A photo accompanying this announcement is available at https://www.globenewswire.com/NewsRoom/AttachmentNg/1f563a2a-9098-4fed-ad90-2c611729067e
