State of the AI industry

About this video

Sarah Friar is the CFO of OpenAI and a seasoned technology executive and investor known for helping scale ambitious companies through periods of rapid growth. Vinod Khosla is a legendary venture investor and co-founder of Sun Microsystems, widely recognized for backing transformative startups and championing bold, long-term innovation.

The emerging signal is not that AI models are getting smarter—it is that businesses are beginning to move from using AI as a chatbot to using AI as a workforce layer composed of agents that execute real business processes. OpenAI's leadership argues that the biggest constraint is no longer model capability but organizational adoption and workflow integration.

For SME operators, the implication is significant: competitive advantage may increasingly come from how effectively a business redesigns workflows around AI agents, rather than simply gaining access to the latest model.

Full Video at the end of page

Core Insight (Plain English)

AI is entering a new phase.

The first phase was asking AI questions. The next phase is assigning AI responsibility for tasks.

Instead of helping someone write an email or summarise a document, AI systems are beginning to manage workflows such as:

• Contract reviews

• Financial reconciliation

• Procurement processes

• Customer service workflows

• Software development

• Planning and coordination tasks

The important shift is that many organisations are still using only a small fraction of available AI capabilities, while early adopters are already redesigning entire business functions around AI-assisted operations.

The assumption being challenged: AI is not primarily becoming a better search engine or chatbot. It is becoming operational infrastructure.

What this means for operators

1. Workflow redesign may matter more than model selection

Many SMEs are still comparing AI models. The stronger opportunity may be identifying repetitive workflows that can be automated or delegated to AI agents.

The question becomes: Which tasks consume staff time but add little strategic value?

2. "People + Agents" may become a common operating model

One example discussed was organisations thinking in terms of one employee supervising multiple agents rather than performing every task themselves.

For SMEs this may mean:

• Leaner administrative teams

• Faster execution

• More capacity without proportional headcount growth

3. Early productivity gains appear concentrated among adopters

The discussion highlights productivity improvements among organisations that have deeply integrated AI into operations rather than using it occasionally.

Operators should watch for a widening gap between:

• Businesses experimenting with AI

• Businesses redesigning processes around AI

4. Vertical specialisation is becoming more valuable

General-purpose AI is improving rapidly.
The opportunity increasingly shifts toward:

• Industry-specific workflows

• Proprietary company data

• Compliance processes

• Operational know-how

This suggests SMEs should focus on combining AI with unique business knowledge rather than trying to compete on AI itself.

5. Adoption may be a bigger challenge than technology

One recurring theme was that most users are utilising only a fraction of available AI capabilities.
The bottleneck is not access to AI, but may increasingly be:

• Training

• Process redesign

• Change management

• Management willingness to experiment

6. Startups and SMEs still have room to build

Despite concerns about large AI providers expanding into more markets, both speakers argue that opportunities remain in:

• Proprietary datasets

• Workflow orchestration

• Governance

• Permission systems

• Industry-specific applications

Particularly where business processes are complex and embedded within organisations.

Practical watchpoints

Watchpoint 1: Agent Adoption Inside Competitors

Monitor whether competitors begin using AI agents for:

• Sales operations

• Customer support

• Finance

• Procurement

• Reporting

The risk is operational speed, not model quality.

Watchpoint 2: Employee-to-Output Ratios

Track whether competitors can grow revenue without proportional hiring.
This may become an early indicator of successful AI integration.

Watchpoint 3: Workflow Software Evolution

Watch ERP, CRM, accounting, procurement and HR systems.
Many are likely to evolve from software tools into AI-managed workflows.

Watchpoint 4: Data Ownership and Permissions

As agents access more systems, governance and permissions become more important.
Poor controls could create operational and compliance risks.

Watchpoint 5: Customer Expectations

Customers increasingly exposed to AI in their personal lives may begin expecting:

• Faster response times

• Personalised recommendations

• Self-service assistance

• 24/7 support

This consumer expectation may spill into B2B markets.

Summary & Reflections

The signal is compelling, but it should be interpreted carefully. Many predictions about widespread agent adoption remain forward-looking rather than proven.

Most companies today are still at an early stage of deployment. Even within the discussion, only a minority of organizations are reportedly using agentic systems extensively.

The practical risk for SMEs is not failing to build advanced AI systems. The greater risk may be:

• Ignoring workflow automation opportunities

• Waiting too long to experiment

• Assuming AI adoption will happen naturally without organizational change
  
  

Regional Consideration (Southeast Asia)

Southeast Asian SMEs often operate with lean teams and fragmented systems. This may actually create an advantage.

Unlike large enterprises burdened by legacy systems, SMEs may be able to redesign processes faster if they focus on targeted operational use cases rather than enterprise-wide transformation.

Who should watch the full video

Most relevant for:

• SME owners

• Startup founders

• Operations managers

• Finance leaders

• Technology decision-makers

• Digital transformation teams

• SaaS founders

• Investors evaluating AI adoption trends

Decision Rating

Decision Usefulness ★★★★★
This signal directly affects staffing, workflow design, software purchasing, and operational strategy. It offers practical implications that SMEs can begin testing immediately.

Operational Relevance ★★★★★
The discussion focuses heavily on finance, procurement, coding, administration, and process automation—areas directly relevant to day-to-day business operations.

Timing Sensitivity ★★★★☆
The shift is already underway but still early. Operators do not need to overhaul their organizations immediately, but delaying experimentation for several years could create a meaningful competitive disadvantage.

Until next time,
The SME Signal editorial Team

Contact

Questions? Reach out anytime.

Email

hello@smesignal.com

© 2026. All rights reserved.