Digital Labor: How AI is Shaping the Future of Business

How jobs get delivered is changing

The rise of AI has ushered in a new era for businesses, providing the opportunity to fundamentally transform operations and creating opportunities for innovation. This shift emphasizes a new conception of AI as ‘digital labor’, where organizations redesign processes to leverage AI’s strengths as a pseudo-employee rather than a simple tool. This is moving beyond simply integrating AI into existing workflows and instead redesigning those workflows for a human-digital hybrid workforce.

That’s one of the key insights from the new World Economic Forum Future of Jobs Report 2025. Based on data from 1,000+ companies, 26 industries, and 14 million workers, the report outlines what’s changing — and what’s coming next. The message is loud and clear: Artificial Intelligence will reshape the labor market faster than most expect.

AI could become a general-purpose technology as transformative as steam engines were in the 19th-century Industrial Revolution. There are clear signs that over the next 2-3 years, AI will flip business and labor upside down.

𝗢𝘁𝗵𝗲𝗿 𝘀𝘁𝗮𝗻𝗱𝗼𝘂𝘁 𝗶𝗻𝘀𝗶𝗴𝗵𝘁𝘀 𝗳𝗿𝗼𝗺 𝘁𝗵𝗲 𝗿𝗲𝗽𝗼𝗿𝘁:

• There will be 170M jobs created and 92M jobs lost by 2030.
• 39% of current skills will be outdated.
• AI is both the #1 disruptor and opportunity.

The view is that digital labor i.e. agents doing work for, with or instead of humans is a $3-12 trillion market opportunity. As the last people to manage an entirely human workforce, we stand at the very beginning of a very disruptive time for business. Every manager in the future will be overseeing hybrid workforces.

Adapting your business to the age of AI means reinventing your conception of what labor is, and reshaping your systems, processes and culture to adapt to that. Even if you don’t, you can be sure that your competitors will.

What are digital workers?

Marc Benioff, the CEO of Salesforce, is a leading voice in the “digital labor movement”. He believes this shift is not just a technological advancement but a fundamental economic transformation, comparing it to the rise of cloud computing and the internet. Some industry analysts believe it could be an even more transformative shift still.

The vision is where humans will work alongside a “limitless workforce” of digital agents, freeing them from repetitive, mundane tasks to focus on more strategic, creative, and high-value work. This movement is not just a theory; Salesforce has been an aggressive adopter of this model, using its own AI platform, Agentforce, to automate a significant portion of its internal work, particularly in customer service and engineering.

While Benioff’s perspective on digital labor is optimistic, it also raises critical questions about the future of employment and the potential for job displacement. He acknowledges that today’s CEOs may be the last to manage exclusively human workforces and that the digital labor revolution will necessitate a fundamental redesign of operational processes, which were designed based on the limitations of humans. The redesign should take advantage of the capabilities of agents.

As Bessemer Venture Partners, put it in their recent report State of AI in 2025: “Don’t bolt AI onto legacy software—reimagine the entire workflow.”

What is an agent?

First, it is worth distinguishing between different types of AI usage in the workplace. There is a huge debate about what is an agent vs AI enabled workflow vs deterministic automation. Whlist not everybody will agree with my definitions, I want to provide them so that the rest of this article can be read in the correct context.

Below are the attributes of these 3 different categories of automation that deliver work. Not all 3 categories are agents. Let’s say that conversational and AI workflow are agents as they are delivered via the Agentforce platform.

Will Salesforce AI agents replace human workers?

While the rise of AI agents is leading to significant transformations in the workplace, the dominant theme is that AI primarily augments human capabilities rather than outright replacing them. This is because few agents are going to be truly autonomous for all but the most straightforward of tasks. Many tasks will require a human in the loop. So agents will be powerful assistants doing the boring, data crunching analysis and offering up recommendations.

In the Salesforce world, which is predominantly front-office processes, AI Agents are unlikely to replace workers. For example, they can efficiently manage daily productivity tasks like writing emails, summarizing documents, and transcribing video conferences. This offers immediate timesaving benefits and facilitates broader AI adoption within an organization. On a more strategic level, AI agents can streamline complex processes, such as conducting due diligence across thousands of M&A documents, identifying critical trends in customer surveys, or analyzing sales calls to coach reps.

Box.com’s internal AI deployments have shown that AI agents can double content output for marketing teams, streamline HR workflows for onboarding, accelerate customer support responses by 50% through internal knowledge hubs, and automate contract analysis in legal departments. The “Case to Bug ” success story at Elements.cloud further exemplifies this, where a conversational AI agent acts as a “virtual technical expert,” dramatically reducing bug resolution time from over 23 days to just 5 days, while concurrently boosting documentation quality.

This broad spectrum of applications underscores how digital assistants empower teams by providing constant, reliable support, allowing human resources to be reallocated to more complex, empathetic, and strategically vital tasks.

Organizations can reshape how teams operate by using agents as assistants that significantly augment human capabilities. The agent’s goal is to accelerate decision-making, surface relevant content, and amplify employee productivity, freeing human teams to concentrate on higher-order thinking, complex relationship-building, and creative innovation. Achieving this requires a redesign of business processes to take advantage of agents, which is something we dive into in more detail in our other articles.

Are businesses already using digital workers?

The number of AI agents in production is low compared with the level of interest and investment that CEO, CIO and CFOs report in surveys. This reflects that most organizations are still in the early experimentation phases. More than 9 out of 10 (96%) of CFOs have an aggressive AI strategy, compared to only 3% in 2020, according to a global survey of 261 CFOs conducted by Salesforce Research. There is a strong shift from cautious spending to a strategic and more aggressive focus on AI for both productivity gains and long-term revenue growth.

A recent McKinsey report found that 80% of companies report using gen AI, yet just as many report no significant bottom-line impact. They believe the heart of this paradox is an imbalance between “horizontal” (enterprise-wide) copilots and chatbots—which have scaled quickly but deliver diffuse, hard-to-measure gains—and more transformative “vertical” (function-specific) use cases—about 90% of which remain stuck in pilot mode.

Salesforce reported that 8,000 customers now have Agentforce licenses with the customers ranging from experimentation phase through to full production with mutiple agents. The move to a consumption-based pricing means that Salesforce only benefits when agents are deployed to production and start consuming credits. Talking to industry analysts we’ve discovered that Elements.cloud is very advanced compared with most organizations.

Industry analysis and consulting firms in their recent reports agree that to benefit from the “vertical” AI agents that McKinsey talked about, organizations need to rethink their operations. This is more than just buying licenses and bolting agents onto existing processes. 

Across industry leaders, the outlook echoing out is similarly changed-focused and active:

AI agents will become embedded in enterprise software, requiring a strategic framework tailored to specific business needs.  

• Gartner, Innovate Business Models Using AI Agents

Agentic AI demands a “new era for process design” as traditional tools and methods are inadequate.

• Forrester, The AI Agent Pivot: A New Era for Process Design

The redesign of workflows has the biggest effect on an organization’s ability to see EBIT impact from its use of gen AI.

• McKinsey, State of AI

AI success requires “revolutionary process redesign” and a “people-first approach,” not just patching AI into old systems. 

• IBM, What is an AI-first company

Drive continuous reinvention: Front-runners continuously optimize, refine and scale AI-driven insights to maintain a competitive edge.  

• Accenture, Front Runners Guide to AI

To unlock the full potential of AI agents, organizations must redesign processes and reimagine business models for agentic AI

• Capgemini, Rise of Agentic AI

A proven case study of digital labor in action

Here is a case study from Elements.cloud internal use of Agentforce. In many ways it is the perfect example as it follows the approach advocated by Salesforce Professional Services.

The example is Case to Bug. It took just 7 days to go from idea to deployment and after 4 weeks the results are staggering

• Avg. Time to Close: 23.8 days to 5.75 days – 76% faster
• % Resolved Bugs: 50% to 100%
• Bug Documentation Quality: 0.94/10 to 8.0 / 10 – massive 751% jump

It starts with the analysis of the business processes to understand the issues and the root causes.  Often, people are too eager to start building agents and jump on the first opportunity. Instead, they should spend time looking at the end-to-end process to work out which agent will have the greatest impact. 

In this case the processes were well documented as UPN diagrams in Elements.cloud. We were able to use the Agent Finder capability that reviews the UPN diagram, and it came up with 5 different opportunities for agents, and it categorized them.  We could pinpoint the agent that would solve the issue, and in this case it was a conversational agent. We used Agent Designer to map out the key steps with instructions and actions. Agent Designer exports the instructions to build the agent in Salesforce Agentforce.

The agent we built reviews a customer Case against the 45 page case analysis document, updates case, identifies gaps in data, and builds an action plan for Customer Support to collate all the information that was missing.

There was a second issue and this involved a different UPN process diagram. Again, Agent Finder highlighted several agent use cases. Again, we used Agent Designer.

The AI Workflow agent that solved the issue reads a Bug record and all related Cases and associated descriptions, emails, notes, and call transcripts. It updates Bug description in the format the Dev team needs, submits a Jira ticket, and allocates it to the correct Dev team.

Megan Higgs designed and built the agents. She is a Business Analyst, with only a year in the Salesforce ecosystem and no technical background. She recorded a video with her internal customer, the Chief Product Officer. The video is a deep dive into the approach, benefits and how they developed the agents:

The entire process can be summarised as:

Forward look

We are in the very early stages of agentification. The maturity and deployment is still very low in most organizations. There are very few established best practices for ideation, build, and governance. There are not many use cases to inspire and inform. The tooling is still in its infancy.

But this will change. The benefits when you get it right are overwhelming. It is not marginal gains in productivity. It is 10x or 100x.