Why the next phase of Salesforce AI will be built around industry expertise

For a while, the enterprise AI conversation seemed to be dominated by a fairly simple question: what can this technology do?

But following Dreamforce 2026, a more useful question for Salesforce leaders is cropping up: what can AI do in our industry, with our processes, customers, regulations and data?

Salesforce’s most recent examples of agentic AI show the technology moving beyond generic assistants and into much more specific operational roles. In higher education, universities are exploring AI agents that support individual students and reduce administrative work.

In financial services, Salesforce is highlighting agentic workflows across banking, insurance and wealth management, while Siemens has demonstrated Agentforce in sales and service.

The pattern matters, because the next phase of Salesforce AI is unlikely to be defined simply by who adopts the most agents. It will be defined by who can give those agents enough business and industry context to make them genuinely useful.

Generic AI can only take businesses so far

Think of the first wave of generative AI as the Swiss Army knife phase. It could summarize a document, draft an email, brainstorm ideas and tackle plenty of other everyday tasks without needing to know much about the organization using it.

Enterprise AI has a tougher job. An AI agent working inside a bank needs to understand a very different environment from one supporting university students or helping an industrial sales team. The workflows, data, risks and measures of success can vary significantly between industries.

That is why the move toward industry-specific Salesforce AI is important. At Dreamforce, Salesforce highlighted financial services use cases spanning origination, service operations, compliance, claims, underwriting and distribution. Its post-event education coverage also showed Penn State and DePaul University using agents to address challenges specific to the student journey.

These are not generic chatbot use cases with a new label attached. They depend on understanding how work actually happens inside a particular organization and sector, and that changes the AI conversation.

Higher education shows what industry-specific AI looks like

Salesforce’s post-Dreamforce look at higher education serves as a useful example. Penn State has more than 87,000 students and 37,000 employees, making personalized support difficult when information is spread across different systems and administrative processes.

The opportunity for AI agents is not simply to answer more questions. It is to use connected data and context to support individual students while taking repetitive administrative work away from employees.

A generic AI assistant might explain a university policy. An agent connected to the right institutional data could potentially understand where a student is in their journey, identify what they need and help move the relevant process forward.

Before you know it, AI needs much more than a good prompt. It needs reliable data, carefully designed processes and an understanding of how the institution operates. In other words, it needs context.

Industry knowledge could become an AI advantage

This creates an interesting consequence for Salesforce teams. As AI becomes more capable, you might assume deep business knowledge becomes less important. The opposite could happen.

Someone still needs to decide what an agent should do and understand which exceptions matter. Someone also needs to know when a process can be automated, when a person should take over and when an AI-generated answer looks technically plausible but is commercially, operationally or legally wrong.

Let’s take insurance as an example. Salesforce’s Dreamforce material highlighted potential agentic applications across claims, underwriting, distribution and policyholder service. Building an agent into those workflows requires more than Salesforce knowledge.

The team needs to understand how a claim progresses, which information matters at each stage, where approvals happen, which decisions carry risk and when human judgment is essential. The same principle applies elsewhere, whether that means understanding the student lifecycle in higher education or complex sales and service processes in manufacturing.

AI may be the engine, but industry knowledge helps provide the steering wheel.

The Salesforce skills conversation is getting more nuanced

This shift has implications for how organizations think about Salesforce skills. The temptation is to look at the rise of Agentforce and conclude that everyone now needs to become an AI specialist.

Some will, but organizations also need people who can connect new AI capabilities to existing business realities. Increasingly, that means combining three areas of expertise: Salesforce platform knowledge, practical AI capability and an understanding of the industry and business processes where the technology will be used.

The sweet spot seems to sit in the space where those areas overlap. For employers, that means a technically impressive Salesforce resumé may not tell the whole story, because experience solving similar business problems can become just as important as familiarity with the latest feature.

Powered by Mason Frank, Tenth Revolution Group works across the Salesforce ecosystem, giving us visibility into how these changing requirements are shaping the capabilities organizations need from Salesforce professionals. As AI adoption becomes more specialized, that combination of technical expertise and relevant industry experience is likely to become increasingly valuable.

AI agents need business processes worth automating

There is another lesson here that is easy to miss. Industry-specific AI is not simply about giving an agent more industry data. Organizations also need to examine the processes underneath it.

If a workflow involves unnecessary approvals, disconnected systems and five spreadsheets named some variation of FINAL_v7_USE_THIS_ONE.xlsx, adding an AI agent does not magically make it a good process. It may just help the bad process move faster.

That is why Salesforce AI transformation needs to begin with the business problem rather than the technology. Leaders should ask where the current process breaks down, what information is required to make good decisions, which steps require human judgment and where an agent could genuinely remove friction.

Those questions are much more useful than starting with: ‘Where can we deploy Agentforce?’ The difference sounds subtle, but it separates AI experimentation from business transformation.

More specialized AI raises the stakes for governance

The deeper AI penetrates industry workflows, the bigger the impact of its actions can become. An agent helping an employee find an internal document carries one level of risk. An agent participating in a regulated financial process, accessing sensitive student information or taking action within a customer account carries another.

Salesforce has placed a lot of emphasis on trusted data, permissions, security and governance as central components of its Agentic Enterprise strategy. Its new AIforce layer, for example, is designed to make Salesforce data, workflows, business logic and governance available to AI interfaces beyond the traditional Salesforce UI.

That matters because industry-specific AI cannot be separated from industry-specific risk. Organizations need people who understand not only what the technology allows, but what the business should allow.

Governance therefore cannot sit solely with the technical team. Legal, compliance, security, data, operations and business stakeholders may all need to contribute to how agents are designed and controlled. The more useful an agent becomes, the more important those conversations become.

Your best AI talent may already understand your industry

There’s something of a workforce opportunity hiding in this shift. Organizations do not necessarily need to look outside the business for every emerging AI capability.

Experienced Salesforce professionals already understand customer journeys, processes, systems and organizational quirks that could take someone new months to learn. Give those folks structured opportunities to develop AI skills, and their existing knowledge becomes even more valuable.

Someone who has spent years working with Salesforce in financial services does not suddenly become obsolete because agents arrive. They could become the person best positioned to explain how those agents should work. The same applies to Admins, Developers, Architects, Consultants and Business Analysts.

Roles will change, but existing expertise does not disappear. It becomes part of the context required to make AI useful.

For leaders, that creates an important workforce question: which AI capabilities should we hire and which should we build within the people who already understand our business? The answer will rarely be entirely one or the other.

The next competitive advantage may be context

Salesforce AI is getting more capable, but access to powerful AI technology is becoming increasingly widespread too. If similar technology is available to everyone, simply having access to AI becomes less of a differentiator.

How an organization applies it becomes more important. That means understanding the customer, the workflow, the data and the industry, then having people capable of bringing all of those elements together.

Dreamforce 2026 showed plenty of impressive technology. The more interesting signal may be where that technology is heading next: AI agents that are less generic and more deeply embedded into specific industries, processes and roles.

For Salesforce leaders, that changes the skills equation. The future may not belong to teams that know the most about AI in isolation, but to those that know how to combine AI with deep Salesforce and industry expertise.

Building the Salesforce capability your AI strategy needs

As Salesforce AI moves deeper into real business processes, finding the right mix of platform, AI and industry expertise becomes increasingly important.

As part of Tenth Revolution Group, Mason Frank helps organizations hire and upskill specialist Salesforce professionals with the technical and business knowledge needed to support transformation, from emerging AI capabilities to established Salesforce programs and beyond.