Sales has long involved a human touch. Where does AI fit in?

Stakeholders and executives are seeking to weave AI into every part of enterprises, and sales teams have not been spared. But in such person-to-person jobs, companies are better off using AI to support employees than to replace them, Andy White, senior vice president of business technology at Salesforce, told CIO Upside.

“We’re very much into this idea of augmenting our humans and having humans and AI working together,” said White. “We can now engage with all these other customers at scale that we never could before because of AI.”

While deals are ideally inked by humans, the sheer scale of an organization like Salesforce means that potential leads often fall through the cracks, he said. “We have millions of leads that we’ve never followed up on,” White added.

With AI agents, those leads no longer go ignored.

• For example, leads are often ranked on a scale of one to five using machine learning, said White, with lower ranking leads getting automated email marketing campaigns rather than human follow-ups.
• With AI, instead of an email campaign coming from a generic no-reply email address, the message comes from a specific, interactive agentic email address, allowing the potential customer to ask questions. The team then measures the success of the agents by tracking email opens, replies and meetings booked with human sellers.
• Salesforce built a new-business pipeline of $500,000 in three weeks by using agents to scour 8,000 previously untouched leads. “It’s pure creation of gold from sawdust on the floor,” said White.

To make agentic email campaigns even better and keep them “from going off the rails,” Salesforce relies on a concept called “coherence” that involves testing the output of one AI model with other AI models. For example, if an email campaign was generated with an OpenAI-powered agent, Anthropic’s Claude and Google Gemini will be used to score it, he said.

“You’re using multiple LLMs that were created using different systems to judge, and if they agree, you have coherence,” said White.

But building these agents isn’t a task for a development team alone, White said. Enterprises need to think about “what is the heart of a seller, how do we give it persistence and grit and not taking no for an answer,” he said. Doing so involves getting the domain expertise from sales teams themselves. “Ideally, you’re working directly with the people that are using the solution,” he said.

“The burden of building great AI and agentic solutions is very heavy on the business teams, and not just the IT teams,” said White. “In traditional technology development, it’s the other way around. But you need to be working directly with the sellers … directly with the people that know what awesome looks like.”

The regulatory landscape around AI may become even more uncertain as Silicon Valley heavyweights invest in curbing the imposition of new rules.

Industry leaders are putting more than $100 million into political-action committees advocating against AI regulation, The Wall Street Journal reported earlier this week. VC firm Andreessen Horowitz and Greg Brockman, president of OpenAI, are helping launch an AI-focused super-PAC network called Leading the Future that will back campaigns against candidates and policies that seek to regulate the technology.

Meta, meanwhile, is preparing to spend tens of millions on its own political arm with a similar focus called Mobilizing Economic Transformation Across California, aiming to support candidates who favor light-touch regulation, according to POLITICO.

“It’s a signal that companies want to be more involved in the AI regulatory conversation,” said Betsy Cooper, executive director for the Aspen Tech Policy Hub “The divide between Silicon Valley and DC is starting to close.”

With investments in political action committees, tech figureheads are throwing their weight behind deregulating AI “in parallel” with their stakes in major model providers, said Thomas Randall, research specialist at Info-Tech Research Group. They’re seeking to preempt the patchwork of state laws starting to take form.

“They’re trying to lock in certain rules … that make building and selling AI easier, cheaper and faster,” said Randall.

However, as enterprises face murky waters with their AI investments and deployments, loosening rules around development could make it even more complicated, said Randall:

• While large model providers could see a shorter time to deployment and less red tape in scaling their AI infrastructure, a “strong majority” of organizations still don’t have strong governance and security protocols in place to internally regulate their use of AI, he noted.
• “If there are limited regulations about implementation of the models within the enterprises, and you have immature organizations trying to leverage these solutions … there will be security pieces that come into play,” he said.
• Additionally, large enterprises that operate internationally may be forced to comply with different rules based on the regions in which they’re doing business, said Randall.

Organizations may begin navigating the regulatory labyrinth by establishing a “governance baseline,” such as following ethical practices created by the National Institute of Standards and Technology. Additionally, being shrewd about copyright and AI output indemnities in vendor contracts could help an organization avoid losing control of its data.

Still, enterprises creating their own regulation may create a headache for the major model providers trying to sell to them, Randall added: “If you don’t have any (regulatory) floor and everyone’s creating something, then you still end up with patchwork – but not across states, just across organizations.”

Oracle wants your data to go with the flow.

The company is seeking to patent a system for provisioning and managing “serverless database resources within a multi-cloud infrastructure,” aiming to make the movement of data and cross-pollination of services between private clouds frictionless.

“Each cloud environment provides a closed ecosystem for its subscribing customers,” Oracle said in the filing. “As a result, a customer of a cloud environment is restricted to using the services offered by that cloud environment.”

Oracle’s system would break down that barrier, allowing a customer to deploy a service from one cloud platform within another. Before approving a request to do so, the system would ensure that the primary private cloud platform has the available computing resources, a data center is close enough to ensure low latency, and can comply with regulatory needs of the customer’s region, such as different state laws.

The cloud service is then deployed into the primary private cloud platform, meaning that the system automatically spins up virtual machines to run the service, and a secure data flow is set up between it and the other cloud platform. The system may also split different workflows between different data center regions that have higher availability.

The goal is to allow cloud customers to get the best of both worlds when it comes to multi-cloud architecture: Rather than having data and services living in individual silos of different cloud environments, Oracle’s tech could offer interoperability.

Patent applications like this from Oracle are no surprise. The company’s primary moneymakers are its cloud services, though it still sits behind the hyperscalers in terms of market share, hovering at around 3% in the most recent quarter, according to CRN.

But with the increasing demand for AI, cloud demand is rapidly skyrocketing. One report from Canalys found that cloud infrastructure spending jumped 21% in the first quarter, reaching $90 billion, largely as a result of AI. Since the rising tide of AI is raising all cloud providers’ ships, tech like this patent describes could make enterprises’ lives easier by making cloud providers play nice with one another.