In a Driving Tomorrow podcast interview with AiNews.com, Cox Business Vice President of Commercial Marketing Sarah Kim argued that AI adoption is not an IT rollout. Companies cannot become AI-first by giving employees AI tools and expecting the existing way of working to produce better results.

For marketing, sales, customer experience, pricing, analytics, IT, product, and executive teams, the decision is how to connect AI tools to customer data, governance, employee readiness, and daily workflows so AI can produce measurable business value.

Kim said many organizations still approach AI as something they can add to existing processes.

“AI is not an IT project,” Kim said. “It is not something that you sprinkle into how you currently do things.”

In short, Kim argues that AI-first growth requires companies to redesign how work gets done. AI can improve customer understanding, automate execution, and support more relevant go-to-market decisions when companies organize data, governance, people, and workflows around it.

An AI-first operating model is the set of processes, responsibilities, data flows, governance, training, and decision points that allow AI to become part of how a company works, instead of a separate tool layer employees use on their own.

AI-first go-to-market redesign means changing how teams use data, workflows, governance, training, and decision-making so AI can support measurable business value.

Kim said many companies treat AI like software that arrives ready to use. That assumption can make leaders underestimate the organizational work required before AI creates repeatable value.

She compared the expectation around AI with software-as-a-service tools that come preconfigured and aligned to known business needs. AI behaves differently because employees can use the same tool in many different ways.

“One of the things that I noticed immediately is you can give the same tool to 100 employees, and they'll use it 100 different ways,” Kim said.

That creates a practical management problem. AI access alone does not tell a company where value is repeatable, which workflows should change, how results should be measured, or how teams should share what they learn.

Kim said companies need a disciplined approach to rollout, usage, and measurement. Without it, AI adoption can spread across many small experiments without producing one high-value business case.

For Cox Business, AI-first adoption began with a roadmap. Kim said she initially asked Eric Pace, who leads the company’s AI center of excellence, for AI because she knew it was important. Pace pushed back and told her the work needed a destination, a full picture, and time to build a roadmap.

“Look, we've gotta have a roadmap,” Kim recalled him saying. “You've gotta understand, the full picture. You've gotta understand the destination that you're trying to get to, and we've gotta spend a couple weeks just doing the work and putting together a roadmap.”

That roadmap gave Cox Business a way to sequence products and prepare employees at the same time. Kim said the product roadmap and the people roadmap had to develop in parallel because AI adoption changes both the systems a company uses and the way employees work.

That parallel planning kept the rollout from becoming a tools-first exercise. Each new capability asks employees to change a habit, hand off a task, or decide where human review belongs, so the organization has to be ready before the product reaches the workflow.

Kim said companies often talk about AI enablement without defining what enablement means at each stage of adoption. For Cox Business, employee adoption began before formal training.

“So first up is it doesn't start with enablement, actually,” Kim said. “It starts with belief. It starts with passion.”

Kim said employees needed to understand that Cox Business could not remain competitive without AI and that marketers also need AI skills to remain competitive in their own careers.

Cox Business then made training directly applicable to employees’ jobs, created communities of practice, paired super users with people who needed more support, monitored usage, and recognized high users as examples for others.

Kim said Cox Business reached full enablement across its marketing organization in about 2 months, covering 100 people.

After that, Cox Business moved into more advanced training. Kim said the next level included building agents, using more advanced tools, and working with n8n. The company’s people-development roadmap mirrored the AI roadmap, so employees built skills as the company’s AI products advanced.

Kim also said leaders need to prepare employees for constant change. AI products may work well for a period of time and then become unnecessary as new capabilities arrive. That can be hard for teams that spent weeks building and testing a product they are proud of.

That pace changes how teams have to think about failure. Kim said employees need room to test ideas, push the limits of the technology, and learn from attempts that do not work yet.

“What they fail at today because the technology is moving forward so quickly, three weeks from now or three months from now it might work,” Kim said. “We're not failing at something, we are not trying hard enough.”

The same lesson applies to products that already work. Kim said leaders may have to tell teams that an AI product they built and like will no longer be needed because the organization has learned enough to move to something better. The product may have served its purpose for one stage of learning, even if the organization later moves on.

Once the AI roadmap defines what will be built and employees are prepared to use AI inside the workflow, the next question is where that operating model creates business value. For Cox Business, Kim said one of the clearest answers is customer understanding, which she described as one of the most important advantages of AI-enabled go-to-market.

Marketing and sales teams often have enough information to see a customer opportunity, but acting on it in real time can still require too many manual steps across systems, data sources, and teams.

Kim said AI can reduce that execution gap by automating parts of go-to-market execution and finding patterns across customer information that would take people much longer to analyze manually. At Cox Business, she said, that has created “a deep understanding of our customers that we've never had before.”

“What we did at Cox is we actually used AI to put together a new segmentation model for our customers,” Kim said.

The company combined about 30 internal and external data sources with 700 attributes, including industry, sub-industry, and customer size. The result was 9 customer segments that showed who customers were, how they bought, and what they bought.

Kim said that structure went beyond a traditional size-and-industry approach.

“It wasn't just size and industry, which is how most companies go to market, if they don't have those deeper insights,” Kim said. “So we used AI to put together a framework and a model for how we segment customers based on how they actually buy.”

Kim said Cox Business was then able to go deeper because AI operates as a platform across the organization, supported by shared data and customer insight.

Cox Business also uses AI to make unstructured customer information more usable. Kim pointed to a tool called Customer Sense, which pulls from sales discussions, customer chats, and service calls so the company can extract customer insights from information that previously sat in less usable formats. She said AI made that kind of customer understanding possible in a way Cox Business had not been able to do before.

Kim said that deeper customer understanding can improve engagement by helping teams act more quickly, automate execution, and reach customers in more relevant and personalized ways.

Kim said marketing can help accelerate AI adoption because it touches many other business functions, including finance, product, industry teams, sales, and customer organizations. That cross-functional role allowed Cox Business to look across go-to-market teams for data, insights, and use cases that could serve more than one function.

The goal was to create shared use cases that could extend from one team to another. One example was content customization.

Cox Business built an AI product for content customization, with an initial use case focused on marketing emails, blogs, white papers, and other marketing content. The company then extended the same knowledge base and workflow to sales follow-up emails.

That knowledge base included brand guidance, legal guidelines, segment messaging, industry messaging, and product information. Because the underlying information applied to both marketing and sales, the workflow could support more than one part of the customer journey.

Kim said Cox Business also found that some of the agents informing marketing strategy could inform sales strategy as well.

“So it's common insights, it's common tools, and it's common AI workflows that power our go to market, and that's been a game changer for us,” Kim said.

The key point: Cox Business is using AI as shared go-to-market infrastructure. A workflow that starts with content customization can support marketing strategy, sales communication, and customer engagement when the same data, messaging, and rules apply across the journey.

Kim said that cross-functional model also changes how companies think about customer management. When teams can work across shared data, insights, and workflows, they can put the customer at the center more directly.

“I really look at our go-to-market is we're not managing different channels, we're not managing different, ways to go to market,” Kim said. “We're actually managing the customer journey and the customer experience with us.”

Kim said that kind of operating model could also change executive roles. She said the CMO role may change significantly as companies organize more work around the full customer journey, potentially creating roles such as a chief customer journey officer responsible for shepherding customer touch points across the business.

Kim said agent sprawl creates business risk because AI amplifies problems that already exist in siloed organizations. When different teams build tools, agents, and workflows without shared ownership, companies can face duplicated work, security issues, inconsistent customer experiences, and unclear accountability.

“It's all of that, right?” Kim said. “It is all of the above and then some, and I think just AI amplifies those issues that we've seen just with having siloed organizations.”

Cox Business uses a centralized AI squad to manage the platform, governance, security, and rollout structure. Kim said that centralized approach has been a foundation for the company’s AI work because teams can focus on the highest-value use cases and the first processes that need to change.

Centralized governance does not mean business teams hand AI work to technology teams and wait. Kim described a co-ownership model where marketing, sales, and the AI squad work closely on product development, testing, adoption, and outcomes.

“Literally my marketing and sales folks and the AI squad are literally hip to hip on a daily basis,” Kim said. “Sitting side by side, developing these products, testing these products, and all of the things. It is truly a co-ownership model.”

That accountability applies to both success and failure. Kim said her team is just as accountable for product outcomes as the AI and technology teams they work with. If a product does not work, the business team may not have provided the right input, set the right expectations, or taken enough ownership of the result.

“If the product is not working, it's not their fault, it's our fault,” Kim said.

Once governance defines who owns the work, companies still have to decide which AI capabilities they should build themselves and which belong at the platform level.

Kim said companies deciding whether to build or buy AI capabilities should start by asking what differentiates the business and what belongs at the platform level.

Cox Business uses vendors for platform needs, general tools, security, and other work where an outside provider can handle the technical foundation. The company builds internally when the capability depends on its own data, process knowledge, customer understanding, and go-to-market approach.

From a marketing and sales perspective, Kim said Cox Business sees differentiation in how it mobilizes segmentation insights for customers and prospects. That includes content creation, enablement, internal data, external data, and processes that are specific to the company.

Cox Business considered vendors for some of that work, but Kim said integrating a vendor solution could have taken as long as building internally. That experience also challenged a major misconception among some leaders. Kim said AI does not replace large platform systems such as ERP or CRM because companies still need structure to build from.

“AI, at least not today, is not gonna replace, your ERP, your CRM, your big platform things,” Kim said. “Because, there has to be, a structure in place. There has to be, certain things that you build from.”

Kim said AI is better understood as a layer that can connect existing systems, fill gaps, add insight, and automate parts of the stack a company already uses.

“It's get the most out of what you have and use AI to fill the gaps,” Kim said.

Kim said Cox Business is experimenting with what she called an “agentified go to market,” where more workflows move from AI-assisted execution toward agentic systems with fewer human touch points.

“Right now, a lot of our workflows are AI assisted,” Kim said. “But there's very much a human in the loop, and honestly, I don't think there necessarily has to be a human loop at every single stop.”

In the next phase of AI adoption, humans may not need to review every step. Kim said the human role changes when automated systems handle more of the workflow. People may touch the process fewer times, but the touch points that remain become more important.

“Even though there's less human touch points in the loops, they become much more important,” Kim said.

That creates a trust and design challenge. Companies need to decide where a person should intervene, how autonomous loops should learn, and how those loops should evolve once they are in use.

Kim said trust must be built carefully. Companies should begin with lower-risk areas, test how autonomous actions work, and then expand once they understand what the system can handle.

For go-to-market leaders, that means the next phase of AI adoption will require more than automation. It will require clear decisions about which steps need human judgment, which steps can run with more autonomy, how teams measure performance, and how employees learn to trust systems without stepping away from accountability.

Q: Why is AI adoption not just an IT rollout?
A: AI adoption is not just an IT rollout because it changes how a company works. Kim said AI-first adoption affects processes, workflows, governance, employee readiness, business ownership, and accountability. Companies should not add AI to existing habits and expect repeatable business value.

Q: What does AI-first go-to-market mean?
A: AI-first go-to-market means using AI to support how marketing, sales, customer experience, pricing, analytics, and product teams understand customers and act on that information. In Kim’s view, AI has to connect to customer data, governance, workflows, and employee training before it can produce measurable business value.

Q: How is Cox Business using AI in go-to-market?
A: Cox Business is using AI to improve customer segmentation, extract insights from sales and service interactions, customize content, support sales follow-up, and connect shared insights across marketing, sales, and customer teams.

Q: How did Cox Business use AI to understand customers?
A: Cox Business used AI to build a customer segmentation model from about 30 internal and external data sources and 700 attributes. Kim said the model helped the company identify 9 customer segments based on who customers are, how they buy, and what they buy.

Q: How can companies avoid agent sprawl?
A: Companies can reduce agent sprawl by giving AI work clear ownership, shared governance, security review, platform standards, and rollout priorities. Cox Business uses a centralized AI squad while keeping business teams directly accountable for product development and outcomes.

Q: Should companies build AI internally or buy it from vendors?
A: Kim said companies should buy platform capabilities and general tools when vendors can provide the foundation. Companies may need to build internally when the capability depends on their own data, customer understanding, processes, or go-to-market execution.

Q: What is agentic go-to-market?
A: Agentic go-to-market is a model where AI systems handle more workflow steps with fewer human touch points. Kim said that does not remove accountability. It makes the remaining human decisions more important because people intervene at fewer moments.

Sarah Kim’s Driving Tomorrow interview with AiNews.com shows how enterprise AI adoption is moving into the daily work of go-to-market teams. At Cox Business, that includes a customer segmentation model built from about 30 internal and external data sources and 700 attributes, plus shared AI workflows for customer insight, content customization, sales follow-up, governance, training, and agentic go-to-market.

Kim’s argument is that AI adoption is a business operating decision, not an IT rollout. Companies need shared data, clear ownership, governed platforms, employee readiness, and workflows that let teams use AI in repeatable ways.

Marketing, sales, customer experience, pricing, analytics, IT, product, and executive teams should pay attention because AI is changing how customer information moves through the business. Insights from sales calls, service interactions, segmentation data, and customer conversations can support marketing messages, sales follow-up, pricing decisions, product planning, support workflows, and customer engagement when teams use the same data, rules, and accountability model.

AI capabilities are changing faster than many companies can redesign their processes. Leaders need a way to test new tools, retire tools that no longer fit, and decide which AI products, agents, and workflows deserve long-term support.

For business leaders, the decision is where AI can create measurable value first and what has to change before the rollout. That includes which capabilities to build, which platform needs to buy, where governance should sit, where humans should stay in the loop, and how employee training should evolve as systems become more autonomous.

In short, AI-first growth requires companies to redesign how work, data, governance, and people fit together. The technology can help teams understand customers, automate execution, and move toward more agentic workflows, but the value depends on whether the organization is ready to change how work gets done.

The companies that benefit most from AI will likely be the ones that make it part of how customers are understood, served, and supported.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing support, AEO/GEO/SEO optimization, image concept development, and editorial structuring support from ChatGPT, an AI assistant. All final editorial decisions, perspectives, and publishing choices were made by Alicia Shapiro.