It’s 1999 all over again: Salesforce puts SA at the centre of agentic AI

In 1999, a Nobel Prize-winning economist said the internet would fizzle out. He was wrong. So are the AI sceptics of today.

Contributed Content

Salesforce Agentforce World Tour 2026
South Africa is at an AI inflection point, and the window for waiting has closed. Image created with TechNation's TN:AI Workflow.

Salesforce (NYSE: CRM) brought Agentforce World Tour to Kyalami Convention Centre in Johannesburg today, where the company made its clearest case yet for agentic enterprise as the tech shift organisations will face in 2026.

The event, themed “Accelerate your digital journey with Salesforce,” drew enterprise leaders, developers, and Salesforce partners from across South Africa and the broader African continent. 

More than 40 expert-led sessions covered agentic AI strategy, with a dedicated hands-on building area giving attendees direct experience in deploying autonomous agents on the Salesforce platform.

Salesforce Agentforce World Tour roundtable

A media roundtable held during the event put three questions directly to Salesforce leadership: where South African enterprises stand on AI adoption, what the agentic enterprise actually requires, and what the skills challenge beneath both questions demands right now. 

Linda Saunders, Country Manager and Senior Director, Solution Engineering Africa, Salesforce, and Ursula Fear, Senior Talent Program Manager, Salesforce, led the discussion. Ana Alonso, Senior Vice President and GM, MENA, Salesforce, opened the session.

Alonso said most of Salesforce’s customers are already using AI, and they are focused on improving how to better service their customers. “This is where the real value lies,” she said. 

The agentic enterprise: intentional by design

Linda Saunders opened her section of the roundtable with a framing that ran through the entire discussion: South Africa is at an inflection point comparable to the arrival of the internet in 1999. 

In 1999, the internet naysayers were loud. Nobel Prize-winning economist Paul Krugman said the internet would fizzle out and amount to nothing (economically).

The naysayers were wrong then, and the organisations waiting for AI to stabilise before committing are making the same bet now.

Saunders also had a very specific message for customers who believe AI means starting from scratch. It does not. 

Every application investment an organisation has already made across sales, service, marketing, and commerce is already available to an agent. It’s a misconception that companies will be starting from zero. 

She said, “If you’ve been building [on Salesforce applications], you have already started your journey to an agentic enterprise because what you’ve built for a human is already completely available to an agent.” 

The platform architecture Saunders presented sits across four systems: context, work, agency, and engagement. Context is the foundation. Agents perform to the quality of the data they draw on, and the richer and more governed that context, the more reliable the output. 

Salesforce’s metadata layer, built once and available across the entire platform, drives the speed and accuracy differentials the company cited: 

  • 16 times faster to deploy than building from scratch, 
  • 75% more accurate agent deployments, 
  • 90% reduction in prompt engineering time.

South Africa’s skills gap problem

Ursula Fear, Senior Talent Program Manager at Salesforce, took the roundtable into the workforce dimension of the agentic shift. 

The headline numbers she presented were drawn from the WEF Chief People Officers Outlook released in May 2026: 

South Africa’s overall unemployment rate sits at 32.7%, with youth unemployment at 61% among 15 to 24 year olds.

Fear said the skills gap between what AI adoption now requires and what formal qualification frameworks currently deliver is already a very active concern. Platform updates run on four-month cycles but South Africa’s National Qualifications Framework operates on five-year timelines. 

This is a gap we cannot close simply by waiting for the system to catch up. “We cannot do this alone. The sweet spot requires sitting with the business, with customers, with partners. We cannot navigate this if everyone is working in isolation,” she explained.

The framework for workforce response sits across four actions she presented as the 4Rs: 

  • Redesign how work gets done, 
  • Reskill people, 
  • Redeploy talent to unlock agility, and 
  • Rebalance human and agent work and monitor outcomes. 

The deeper question Saunders and Fear kept returning to was not whether AI would change how work gets done, but how far that change goes. Fear presented Salesforce’s Human-AI Collaboration Model as a way of mapping that journey honestly. 

It starts with AI as a tool, something a person commands to complete a single task. First it moves through AI as an assistant, then as a catalyst, and eventually as a contributor working alongside humans on complex problems. 

Fear said each stage in this progression requires a deliberate decision to go further.

ABSA partnership and AWS Marketplace

Saunders announced two significant developments for the South African market at the event. 

The first was an expanded strategic collaboration with ABSA, one of the first banks anywhere in the world to go live with Agentforce back in June 2025. The expanded partnership supports ABSA’s ongoing work to embed agentic capability across its business divisions.

The second announcement was the opening of AWS Marketplace as a purchasing channel for South African Salesforce customers. The decision was driven by customer demand to consolidate cloud spend. 

If the roundtable had a single takeaway, it was this: the window is open and the tools already exist. The South African market is ready and organisations moving on this now will set the pace. The ones waiting for a more convenient moment are, as Saunders put it at the May roundtable, “already behind.”

NOW READ: The window for cautious AI adoption has closed, says Salesforce

Before you @ us:

No, AI did not “write this article.” Calm down. This piece was produced using our TN:AI newsroom workflow. The opinions and typos belong to a human who has algorithmic side quests. (Hi!) We even wrote an AI policy so nobody panics.

🧠 AI-assisted research + summarisation 📝 Human edited + fact-checked

Sharing is caring! 

Featured reads: