Generative AI

IT teams at financial companies have their hands full untangling years, and sometimes decades, of tech debt. Reliance on legacy technology can slow down innovation and put broader business goals out of reach.

AI can step in to speed up the modernization process, according to Roman Eisenberg, head of technology for Chase card and connected commerce at JPMorgan Chase. 

“The core of credit card processing served us well for many decades because it was built for scale and extendability," Eisenberg told CIO Dive. "But it was built on mainframe, using legacy programming languages.”

Over time, however, the mainframe infrastructure made it harder to innovate. Chase was also grappling with skills gaps, making it difficult to maintain let alone expand capabilities.

While initially skeptical about including AI as part of platform modernization efforts, Eisenberg and his team were able to allay concerns through robust governance and guardrails. 

“At first, the criticality, the risk associated with making mistakes and all kinds of hallucinations that we saw, there was no way we could allow it to be anywhere near mission critical [platforms],” Eisenberg said. 

Other enterprise leaders can relate. AI adoption that isn’t supported with adequate safeguards and processes can degrade outcomes. Analysts have urged technology tech chiefs to approach the opportunity with care. 

The potential upsides motivate CIOs to keep trying.

AI-assisted modernization can help organizations unshackle themselves from outdated systems at a faster pace. Around 80% of IT and business leaders expect AI coding assistants to help document legacy applications, rewrite old code and automate software testing, according to a Publicis Sapient report published in May. 

Identifying governance gaps and establishing the right safeguards changed everything for Chase.

“We saw immediately that it allowed us to unlock real value for our customers,” Eisenberg said. “For me, it was one of those 'ah-ha' moments.”

Standing up guardrails

The stakes are high for modernization projects, but success is hard to come by. 

The majority of legacy tech upgrades fail initially, and more than half contend with the stalling of broader digital transformation initiatives, according to a Forrester survey. Mainframe skills gaps, complex integration strategies and inadequate tools are all to blame for misfires. 

“We cannot make mistakes,” Eisenberg said. “You have to move fast, but you absolutely have to minimize any chance of breaking things because it’s how people buy groceries and pay for childcare.”

With that in mind, Chase ensures upgrades occur in a sandbox before moving into production. With tens of millions of lines of code to refactor, teams go after the highest-value modules first. 

“A lot of it pertains to customer experience, so we would prioritize that to narrow it down,” Eisenberg said. “We then create that test ecosystem and apply generative AI tools to explain what the code does.”

From there, teams use the documentation to generate new code in modern languages that are tested, compared and validated. 

“If we see that the same tests work, we have a very high degree of confidence that we didn’t miss anything and that the generative AI didn’t hallucinate,” Eisenberg said. 

Humans in the loop are also responsible for ensuring that newly generated code is up to par. 

“The new code might have to be refactored further or moved to a different domain,” Eisenberg said. “But these are several steps that decide whether it’s ready to be shipped into production.”

Chase card teams have been using the process for about a year and are likely to continue into the foreseeable future. 

“It will end when we decommission the last line of mainframe code,” Eisenberg said. “But we’re not going to stop at mainframe. We have so much other infrastructure where this approach is beneficial.”

As the financial sector harnesses generative AI’s capacity to speed and simplify processes, Bank of America has leaned on sustained research and technology investments to rapidly scale multiple use cases.

The firm unveiled a generative AI-powered knowledge management assistant called AskGPS for its Global Payments Solutions division in October. In-house engineers fed the model 3,200 internal documents, from product guides and term sheets to FAQs. The tool provides tailored answers to the division’s more than 40,000 business clients, according to the Sept. 30 announcement.

“Pre-generative AI, we had standard library search to find documents, not answers, and they simply didn’t have the power of generative AI,” Jarrett Bruhn, head of data and AI for GPS at Bank of America, told CIO Dive. “Deep, difficult questions would take up to an hour of searching.”

AskGPS speaks 29 languages, which is a major advantage for a globally dispersed banking unit, and is already easing the onboarding process for new hires, Bruhn added.

“This is their go-to for acronyms and product knowledge,” Bruhn said. “If you’re in a meeting and supply chain finance comes up or someone mentions a new acronym, you no longer have to spend all that time digging through everything to access the product knowledge.”

The idea for AskGPS began germinating early in the evolution of commercial-grade large language model capabilities. Engineers on the GPS team worked with the bank’s global technology team on the building process, from initial design through deployment.

Collaboration ensured that the tool would be scalable, secure and aligned with the bank’s broader AI strategy to “build adaptable tools that benefit multiple lines of business,” a Bank of America spokesperson told CIO Dive.

AI across banking

Banking provided fertile soil for data-hungry generative AI applications. Bank of America and its peers in the heavily regulated financial industry were well positioned to impose sturdy guardrails around sensitive data amid the proliferation of potential use cases.

“The banking sector is a blueprint for how AI should be rolled out in any sector,” Alexandra Mousavizadeh, co-founder and co-CEO of Evident Insights, told CIO Dive. “It’s already got all of the checks and balances and governance and risk functions you need.”

Bank of America has seen its profile rise in the industry, according to Evident’s annual banking AI index published in October. The bank ranked among the Top 10 for the first time since Evident began tracking AI maturity among 50 of the largest global financial firms in 2023.

AI research has fueled Bank of America’s rise, Mousavizadeh said. The firm’s commitment to innovation has yielded 7,400 patents and pending applications, including 1,200 tied to AI and machine learning.

Development and adoption have been moving along steadily too.

The company saw uptake of its Erica for Employees virtual assistant spread to more than 90% of its 213,000 employees in April. The tool, which was introduced in 2020, has since cut the number of IT service calls by half, Bank of America said.

The employee assistant built on the success of pre-generative AI, consumer-facing chatbot Erica, which was introduced in 2018 and averaged 58 million customer interactions per month as of August. More recently, the company deployed a pair of AI knowledge management tools to its wealth division and saw 17,000 of its programmers incorporate AI coding tools into their workflows.

“Generative AI has helped us do some foundational things both with the outside world and within the bank, where we can accelerate time-to-market and make processes easier and faster,” Bank of America executive Hari Gopalkrishnan told CIO Dive in April, prior to his July appointment as chief technology and information officer.

Gopalkrishnan previously served as the bank’s head of consumer, business and wealth management technology and is now leading efforts to expand the company's AI portfolio, CEO Brian Moynihan said in a memo announcing the appointment.

The AskGPS initiative grew out of broader efforts to improve the banking experience for advisors and their clients, according to Bruhn.

“At the genesis of AskGPS … it was about really figuring out our key deliverable, which is informed advisory and deepening those client relationships,” said Bruhn.

Industries across the economy are continually refining their AI strategies quarter after quarter, often tying the success of initiatives to broader business goals. 

While generative and agentic AI are dominating conversations, organizations are still relying on foundational elements, such as machine learning and routine automation. However, newer technologies have captured CIOs' attention.

Here’s how retailers, banks and beverage makers are approaching AI adoption, from ROI to use cases. 

Retailers

Retail businesses are adopting AI for a plethora of reasons. Some want to use the technology to revamp customer or employee experiences. Others see the technology as a way to accelerate growth or cut costs. Most would say all of the above. 

AI and technology are important levers in Dollar Tree's turnaround plan. Executives described the strategy as a “culture shift,” during its investor day in October. 

“To make the most of our opportunities, Dollar Tree must evolve … from underinvestment in technology to an AI-enabled enterprise,” CEO and Director Michael Creedon said during the presentation to investors. 

As part of its strategy overhaul, the company is replacing decades-old legacy systems with modern cloud-based platforms that can improve assortment planning, inventory visibility, workforce management and its cost structure at scale. 

The company named Travis Nixon its SVP of AI optimization on Tuesday to complement the efforts. 

The Tractor Supply Company is also optimistic about the potential for the technology. The company set three AI priorities: off-the-shelf software; custom-built software; and agents and automation.

Over the last half year, the company has ramped up adoption across teams, executives said. Around 1,500 of its employees have received access to OpenAI’s technology via enterprise accounts that are integrated with its Snowflake Data Lake, according to Hal Lawton, president, CEO and director of Tractor Supply. 

“What that allows us to do now is to start, across the organization, building agents to automate and make things simpler and faster,” Lawton said during the company’s Q3 2025 earnings call in October. 

The company is also working closely with existing vendors to adopt AI add-ons and capabilities where appropriate, “with clarity of understanding of functionality and security,” Lawton added. 

Tractor Supply’s third-quarter results were in line with expectations as AI adoption progressed. Its net sales increased 7.2% to a record $3.7 billion while comparable store sales increased 3.9%.

Beverage makers

Food and beverage companies are betting on AI as a cost-saving measure, but the technology also requires significant investment to get started. 

Molson Coors embarked on a turnaround strategy that includes revamping the organizational structure and pursuing growth opportunities. Even in challenging times, AI investment is a priority. 

Continuing to ensure the company is investing sufficiently in technology is important, Rahul Goyal, CEO, president and director of the beer manufacturer said, pointing to infrastructure needs and AI.

“We're going to continue focusing on these areas… We need to make sure that the investment we have in people, technology is driving the right returns,” Goyal said during the company’s Q3 2025 earnings call.

The Coca-Cola Company is also working to find the right balance between investment and cost savings. The company poured $1.1 billion into expanding its Microsoft partnership last year as it looked to explore Azure OpenAI Service and Copilot. 

“We're going to find the extra investments to drive that growth,” CEO and Chairman James Quincey said during the company’s Q3 2025 earnings call in October. The company is currently assessing how it needs to evolve to generate its desired outcomes.

“That [evolution] will include ongoing productivity as we bring in AI and agentic tech over the coming years,” Quincey said, adding that the company would undergo some restructuring in 2026.

“While the overall environment has continued to be challenging, we’ve stayed flexible — adapting plans where needed and investing for growth,” Quincey said in a release accompanying the financial results for the quarter. 

Coca-Cola reaffirmed its full-year guidance as net revenues grew 5% year over year to $12.5 billion for the three-month period ending Sept. 26. 

As generative and agentic AI become more critical to Coca-Cola’s operations, the company is seeking returns on its investments. 

“The role of AI, the role of generative AI, it's not just interesting to talk about,” John Murphy, Coca-Cola president and CFO, said during a conference this summer. "You've got to now convert that into real tangible outputs.” 

Financial services firms

Banking institutions are chasing AI-fueled efficiencies — and there’s no signs of a slowdown. 

BNY is undergoing transformation as it debuts a new commercial model and embraces AI. Over the past three months, the company has been particularly busy. 

As part of BNY’s third-quarter earnings report, CEO Robin Vince said the company launched its AI platform, called Eliza, in September, democratizing access across its business. During the last quarter, the company has also begun collaborating with Carnegie Mellon University to create an AI lab that advances research, responsible governance and deployment. 

“The investments we've made have been focused on creating the technology foundation to go faster, but adoption and success are ultimately driven by culture,” Vince said during the earnings call in October. “By putting AI in the hands of everyone at BNY, we intend to develop fluency and create capacity for our people to focus on higher-value work.”

At the end of Q3, BNY had 117 AI solutions in production, an increase of 75% compared to the prior quarter. The sum represents AI agents that identify business leads, write code and accelerate client onboarding. 

Other financial services firms are pursuing agentic AI adoption, too. Citi, for example, piloted the technology for about 5,000 colleagues as part of a project that launched in September. 

Agentic AI is set to be a prominent theme for the sector as businesses enter the new year.

“We believe our AI opportunity is significant,” Vince said. “We are pursuing it with urgency.” 

Boardrooms haven't caught up with enterprise efforts to infuse AI into every corporate process, according to PwC data. The firm surveyed more than 600 public company directors for the report.

Nearly 2 in 5 directors believe their boards have not received enough education on AI despite widespread adoption initiatives. Just over one-third of directors said their boards have incorporated AI and generative AI into oversight roles.

“Boards are not alone, the rise of generative and agentic AI means both management and boards are learning about new capabilities and risks in real time,” said Ray Garcia, leader of PwC’s Governance Insights Center, in an email. “Directors aren’t expected to be technologists, but they are expected to understand the technology's potential, limitations and risks so they can ask informed, strategic questions and challenge management when needed.”

The C-suite was thrust into fast-paced change as generative AI landed in the enterprise and expanded into software offerings. Adoption became a top priority as businesses eyed productivity gains.  

But AI is top of mind for board members too. Nearly two-thirds of directors identified AI as the leading priority for their businesses, according to a Diligent Institute report published last month. Almost one-third of directors said they sought AI education in order to get up to speed.

“As companies continue to adopt AI into their business strategy and operations, boards want to better understand its impact and engage with management on the right questions,” said Garcia.

According to Garcia, some of the key questions board members ask include:

• How will AI affect business strategy and customer interactions? 
• What risks does AI introduce and how can the company incorporate responsible AI? 
• How will AI reshape the workforce and talent strategy?

Despite ongoing deployment plans, more than 2 in 5 directors say keeping up with the pace of change is a top concern related to AI, according to the PwC report.

CIOs can leverage their own AI savviness to help expand awareness among top executives. One example comes from Sathish Muthukrishnan, chief information, data and digital officer at Ally Financial, who planned a day of AI training for the company's executive committee in order to increase visibility into adoption efforts.  

“CIOs have an important role to play in helping boards close the knowledge gap on AI,” Garcia said. “They can organize briefings with the board offering insights on LLMs, agents, responsible use, frameworks and other topics. They can also contribute through cross-functional councils that assess AI opportunities and risks across the business.”

Board members should also reach out to other C-suite leaders — including those in legal or risk functions — and seek external upskilling opportunities to build up AI expertise, Garcia said.

Almost three years into the enterprise generative AI craze, enterprises report adoption roadblocks that are difficult to sidestep, according to an ABBYY report. Opinium Research conducted the survey of 1,200 senior managers and other top leaders from the U.S., UK, Germany, France, Australia and Singapore.

About 1 in 3 businesses say training generative AI models is harder than expected, and slightly fewer respondents said staff members don’t have the necessary skills to deploy the technology. Companies are also having a hard time integrating generative AI into business processes and creating proper governance structures. 

Even after implementing generative AI, success isn’t guaranteed. More than 1 in 5 survey takers say employees are misusing the tools, and the same percentage of respondents say employees bring generative AI tools outside IT's purview for personal productivity, raising security concerns.   

While the conversation has shifted to other AI technologies, such as agents, enterprises continue to struggle to overcome integration, training and governance challenges tied to generative AI.

“Businesses spent money on Gen AI tools that promised more than they can provide,” Maxime Vermeir, senior director of AI at ABBYY, said in a blog post. “In some cases, they didn’t even need it.”

One of the tasks CIOs have taken up is sifting through vendor hype and determining where generative AI fits into enterprise workflows. Businesses are motivated by the potential for productivity benefits, according to the ABBYY survey, but they’re also spending more conservatively. More than 2 in 5 survey takers plan to increase their spending by up to only 15% in the next 12 months, the report found.

“A lot of enterprises are more mindful and are taking a harder look at AI investments,” said Eden Zoller, chief analyst of applied AI at Omdia. “You’ve really got to get quite strategic and quite disciplined with the approach.”

Enterprises have had success with generative AI by honing project prioritization processes and addressing other kinks head-on. 

Marriott Global CIO Naveen Manga characterized the hotel giant’s approach to generative AI as building on its foundation of “ruthless” prioritization. Kraft Heinz has put an emphasis on pressure testing use cases before going all-in to prevent dead-end generative AI projects. PepsiCo has also improved processes to better enable generative AI projects to move into production by culling priorities.

Over the next three years, 42% of firms expect to spend more than 30% of their cloud budget on generative AI, according to a survey of 215 North American IT buyers published by TD Cowen in October. Nine in 10 leaders expect generative AI use cases to account for 10% of the cloud budget. 

Enterprise generative AI workloads are moving rapidly to public cloud hyperscalers such as Oracle, Microsoft, Google and Amazon. Generative AI usage has increased across nearly all functions within organizations, according to TD Cowen. The survey findings revealed that IT buyers’ cloud spend on generative AI workloads will rise fourfold by 2028. 

“As more than 90% of large enterprises deploy Gen AI applications by 2028, IT leaders must prepare for higher costs tied to advanced infrastructure and new workloads,” Hardeep Singh, a principal analyst at Gartner, told CIO Dive.

As generative AI workloads increase cloud budgets, it could affect the way IT leaders manage their cloud platform costs. 

CIOs are often taking a multi-pronged approach to managing spend, including negotiating commitment-based contracts with vendors to increase cost transparency and capacity guarantees, Singh said. Deployment of cost management tools also helps IT leaders track usage and optimize available resources. CIOs are increasingly turning to cost tracking practices such as FinOps to analyze cloud spend.  

A key strategy includes “right-sizing cloud environments” by matching resource allocation with workload requirements and implementing autoscale strategies to adjust capacity based on demand, Singh said. 

“These measures help organizations ensure efficient consumption and keep cloud spending aligned with financial goals as AI adoption accelerates,” he said. 

As IT executives seek more value from their cloud products, Singh said enterprises want to go beyond basic cloud services to products tailored to their business needs. Hyperscalers are responding by investing billions into their cloud and AI buildouts. 

Data sovereignty regulatory requirements are also driving demand for specialized or hybrid cloud products to ensure compliance, he added. SAP, for example, is pouring billions into building out its sovereign cloud offerings in Europe.  

“Enterprises are moving beyond generic cloud services toward more industry-specific solutions that address unique regulatory and operational needs,” Singh said.

The rise of generative AI tools will more deeply transform tech jobs more than other sectors, according to an Indeed analysis. The job site's Hiring Lab analyzed more than 2,800 skills for the report to identify what skills would be most and least affected by generative AI adoption.

More than half of the skills that comprise technology roles will undergo deeper transformation, according to Indeed. Nearly 3 in 5 of the job skills that could potentially be fully transformed are related to technology. 

Despite the possible upheaval to technology roles, Indeed cautions that widescale job replacement is not yet in the picture. It found that less than 1% of skills are rated as likely to be replaced by generative AI.

The influx of generative AI skills is shifting how IT departments operate. Although specific tasks are headed toward automation, IT workers remain essential to organizations.

"A job is made up of skills, and often multiple skills," Cory Stahle, senior economist at Indeed's Hiring Lab, told CIO Dive. "Tech jobs have more of the skills that have potential to be transformed, but many skills have at least some human element to them that is not going to be able to be replaced by generative AI."

AI's ability to take over repeatable, rules-based processes helps explain why most high-impact tasks are technology related. Skills mentioned in 82% of software development job postings are among those facing deeper change, according to Indeed. 

"There are set-in-stone syntax requirements when you talk about coding," said Stahle. "The skills where you're processing structured data or you're doing something in a very structured way, generative AI is pretty good at doing that."

Software development teams have been plugging AI into processes for several years as AI's coding chops have improved and vendors have widened their offerings. AI has gained ground in code reviews, prototyping and other repeatable tasks in software development.

Executives are spurring adoption across the enterprise, lured by the potential for improvements to productivity and revenue. But concerns about the accuracy of outputs and the potential effects on career development are weighing down efforts.

As deployments progress, executives should not discount the value of human skills within their ranks, Stahle said.

"The focus of using AI should be to solve problems and to make better products and services for people," Stahle said. "Not to use AI for the sake of using AI."

Target is moving full steam ahead on a turnaround plan with the goal to reverse declining sales and traffic numbers — and technology is playing a key role.

Current COO and incoming CEO Michael Fiddelke outlined his key turnaround priorities during a Q2 earnings call. The executive emphasized the need to regain merchandising authority and said the company must better utilize “technology to improve our speed, guest experience and efficiency throughout the business.”

Now, the company is providing tangible examples, showing how it is using technology — particularly AI — as Fiddelke steps into the spotlight.

From merchandise development to its third-party marketplace, Target is weaving AI across its offerings with the aim to help human creativity along the way. 

That's why Target's Chief Information and Product Officer Prat Vemana believes it's the best time to be working in technology and retail, the executive told Retail Dive in an interview.

Moving fast on trends

The retailer is turning to its generative AI-powered trend intelligence platform, Target Trend Brain, to surface new ideas for its merchants, the company said in a blog post. Machine intelligence is helping to speed up product development in response to trends.

It connects back to Fiddelke’s merchandising authority focus through design, Vemana told Retail Dive.

“What differentiates us is that human element, that we are good in design,” he said.

Target Trend Brain aims to combine merchant creativity with AI-driven efficiency. Target is in the initial phases of using the technology to help the company’s merchandising ideation and plans to create collections at a faster pace, Vemana said.

“When we look at our designers, they're world class,” Vemana said. Such tools are meant to empower these employees’ creativity with the agility of AI. 

“It’s the designers who are the ones making the decision,” Vemana added. “AI is there to help.”

Reviewing third-party sellers

The retailer is taking a similar approach in how it reviews applications from sellers for its marketplace, Target Plus. The third-party seller offering launched in 2019 with the intent of being a more curated version of a retailer marketplace. At the time, Target described the product selection as “thoughtfully selected additions.”

Now, that seller selection process at Target is getting some support from agentic AI.

“We have agents that allow you to analyze every applicant to make sure that the vendor who applied is the right vendor for Target and for our guests,” Vemana said. “Agents allow you to actually go and look at the information available online and brings everything in a summarized form for the marketplace analyst so that they can actually review it in one place.” 

Target is using AI for the upcoming holiday season, too. Vemana said that the retailer rolled out its demand forecasting engine across categories this year, and its predictions improved with every new iteration.

“We are actually confident leaning into the holidays that leveraging the AI for positioning and placement of the products in the right places is going to actually reduce, significantly, any frustration,” Vemana said.

Of course, the widespread adoption of new technologies at a company of Target’s scale requires both employee training and encouragement to explore the tools within the safety of enterprise systems.

Growing workforce skills

In August, Fiddelke told analysts that the retailer had deployed more than 10,000 new AI licenses across its company since its Q1 earnings call in May. The increase in licenses is part of an effort to develop the workforce’s knowledge of AI, according to Vemana, who added that Target even had OpenAI come in to train staff.

While AI in general has been disruptive, the technology isn’t entirely foreign to the company, Vemana said. Target had already used it in its loyalty program.

“What we have done recently is lean into generative AI,” he said. “How do we bring the knowledge of the workforce up and running so that they can use tools like ChatGPT for them to improve their productivity?”

OpenAI has been making its way into the retail space for some time now, changing the search behaviors of consumers who are asking more complex queries to find product recommendations — a topic which Target executives have spoken about before.

But OpenAI has taken it a step further in recent weeks with the launch of Instant Checkout for ChatGPT, which allows users to make purchases in the chat platform without needing to leave it. The feature was launched first with Etsy and Shopify. Now Walmart will also enable the functionality and plans to develop employee training with OpenAI.

In October, OpenAI announced a set of apps users can talk with directly within ChatGPT with partners including Spotify and Zillow. For example, when talking with ChatGPT about looking for a new house, the platform can surface the Zillow app inside the chat and show listings for users to browse with an interactive map. OpenAI listed Target as one of 11 additional partners it plans to welcome with apps later this year.

In general, staying flexible with the types of AI models that Target experiments with — an approach Vemana describes as model-agnostic — is also key for employees like software developers, who are using what the retailer calls ThinkTank to test new ideas.

“We will be able to switch the models very quickly under ThinkTank so that you're not stuck with, like, one provider and one approach to how you problem solve that way,” Vemana added.