Embracing the Era of the Agentic Enterprise

The transition from generative to agentic AI represents one of modern-day computing’s more significant shifts. While the previous era was defined by models that could summarise text or generate imagery on demand, the next age is one of autonomous action.

Agentic systems do not just provide answers. Instead, they execute workflows, reason through multi-step problems and interact with external tools to achieve complex goals with minimal human intervention.

AI, as has been widely asserted, is moving from a consultancy role to that of a co-worker. The scale of the transformation will surely be reflected in the market, with Fortune Business Insights projecting that agentic AI’s value will jump from US$7.29bn in 2025 to an eye-watering US$140bn by 2034. 

Unlike static tools, agents can navigate ambiguity, correct their own errors and manage projects from inception to completion. Research suggests this autonomy is already yielding tangible results: 82% of organisations plan to integrate AI agents within the next three years to handle tasks requiring reasoning and multi-step execution, according to Capgemini. 

For businesses, the implications are profound. Gartner predicts that, by 2028, agentic AI will manage 68% of all customer service interactions, a reality that goes way beyond basic chatbots. 

Furthermore, early adopters in the enterprise space are already reporting productivity gains, with McKinsey noting that 39% of surveyed companies are already running 10 or more agents in production environments.

However, the rise of agency brings a new set of challenges. As these systems gain the power to act on our behalf, focus is already switching from prompt engineering to robust governance, safety frameworks and reliable alignment. 

At Salesforce, Slackbot is evolving into an agentic work operating system that connects large language models to a user’s lived context – from Slack threads and calendars to CRM records and shared drives. 

Rather than waiting for instructions, it acts as a proactive teammate, reducing the ‘toggle tax’ of moving between applications. Employees use it to prepare for cross-functional meetings or to automate entire project lifecycles through multi-agent collaboration. 

Internal research shows 73% of adoption has been driven organically through colleagues sharing effective prompts, signalling a behavioural shift in how work is coordinated. The next step is interface evolution: using Slack’s Block Kit framework to dynamically generate bespoke forms and visual components, unifying search and execution in a single entry point.

Meanwhile, agentic AI is underpinning an AI-first reinvention of customer experience at Zoom. 

The Zoom Virtual Agent moves beyond scripted chatbots to deliver autonomous, end-to-end resolution across voice and chat, retrieving order details, updating records and anticipating needs before they are fully articulated. Crucially, it orchestrates intelligent hand-offs into Zoom Contact Center when empathy or specialist judgement is required. 

This hybrid autonomy reflects wider industry forecasts that agentic systems will handle the majority of routine service interactions within the decade.

Elsewhere, Saudi-based HUMAIN is embedding agentic AI at operating system level. 

HUMAIN OS employs an intent-driven architecture capable of orchestrating complex enterprise workflows, with more than 150 agents already managing functions from payroll to project delivery. 

Backed by sovereign-scale compute and partnerships with leading AI hardware providers, the company is pairing diversified infrastructure with proprietary frontier models tailored to Saudi culture – signalling how agentic AI is becoming foundational to national as well as corporate strategy.

Salesforce’s personal agent for work

Slackbot from Salesforce is evolving from a standard productivity tool into a sophisticated agentic work operating system. By connecting large language models to a user’s specific context – including Slack threads, calendars, Salesforce records and Google Drive – it functions as a proactive teammate rather than a static assistant. 

Unlike traditional AI that requires constant orientation, Slackbot already understands how an individual works and communicates.

This integration eliminates the "toggle tax" of switching between tools. Use cases range from grounding staff before cross-functional meetings to automating the entire lifecycle of a project, such as generating landing pages via a multi-agent collaborative process. Salesforce’s own research indicates that this shift transforms how people prepare for interactions, with 73% of internal adoption driven by the social sharing of successful prompts.

Looking ahead, it is hoped Slackbot will transcend purely conversational interfaces. Utilising Slack’s design framework, Block Kit, it is set to dynamically generate custom UI components, such as forms or visual displays, tailored to specific user intents. This foundation ensures that search and assistance converge into a single entry point, streamlining the “last mile” of workplace productivity and fundamentally altering the relationship between humans and digital agents.

Zoom's agentic AI for customer-centric service

As it evolves into an AI-first enterprise platform, Zoom is embedding agentic AI at the heart of its customer experience strategy. Moving beyond reactive chatbots, agentic AI enables autonomous, end-to-end resolution — managing complex interactions from initiation through to completion without constant human oversight. 

This aligns with projections from Gartner analysts, who predict that, by 2029, agentic AI will autonomously handle up to 80% of routine customer service requests, reducing operational costs by as much as 30%.

At the centre of the shift at Zoom is the Zoom Virtual Agent. By orchestrating workflows across self-service and live environments, the system ensures intelligent handoffs when human empathy or expertise is required.

Ben Neo, Head of Zoom Contact Centre and CX Sales EMEA, explains: “Agentic AI is transforming how Zoom delivers end-to-end autonomy in customer experience by enabling intelligent systems to autonomously manage complex customer interactions, from initiation to resolution.

“Zoom Virtual Agent now autonomously completes complex tasks like retrieving order details or updating personal information using either voice or chat. It anticipates needs, offers solutions before problems are fully expressed and integrates seamlessly with Zoom Contact Center to ensure smooth, context-rich handoffs to live agents when needed.”

HUMAIN’s bold leap into agentic AI

HUMAIN is redefining interaction between machines and human workers with HUMAIN OS. 

Unveiled by CEO Tareq Amin at the 2026 PIF Private Sector Forum, the agentic operating system moves beyond the static environments of legacy platforms like Windows or MacOS. It employs an AI-native, intent-driven architecture designed to orchestrate complex tasks across enterprise systems autonomously.

HUMAIN is already leading by example in the realm of agentic AI. Through its HUMAIN One interface, more than 150 AI agents are able to manage corporate functions ranging from payroll to project management. The strategy directly addresses the 90-95% failure rate often seen in enterprise AI pilots by embedding intelligence into workflows with measurable outcomes. 

“We decided not to take the old path,” Tareq noted during the unveiling. “We decided to reinvent.”

Supported by HUMAIN Core, a network of gigawatt-scale data centres, not to mention strategic partnerships with industry leaders like NVIDIA and xAI, HUMAIN is helping to position Saudi Arabia as a major exporter of compute.

The company follows a diversified hardware approach while creating proprietary frontier models specifically designed for Saudi culture, language and context. This reduces reliance on open-source derivatives and ensures models are optimised to meet local business and societal requirements.