AI adoption accelerates as enterprise PoCs display productivity gains

Like other CIOs, Katrina Redmond has been inundated with opportunities to deploy AI that promise to accelerate business and operations processes, and optimize workflows. “Everyone seems to be scrambling to figure out this technology that’s moving so fast, but without business outcomes, there’s no point in it,” says Redmond, CIO at power management systems manufacturer Eaton Corp. “We need to stay focused on business outcomes and prioritize use cases that make sense.”

Some prospective initiatives require custom model development using large language models (LLMs), while others simply require flipping a switch to enable new AI capabilities in enterprise systems. “AI is showing up in every software package and every technology, especially generative AI,” says Dan Diasio, global AI consulting leader at EY, while some vendors, such as Microsoft, have made AI central to their offerings.

To keep pace, Redmond formed a steering committee to identify opportunities aligned with business goals, and narrowed a long list of prospective initiatives down to a few dozen that range from inventory and supply chain management to sales forecasting. “We don’t want to just chase the next shiny object,” she says. “We want to stay disciplined and go deep.”

To achieve success, an AI proof of concept (PoC) project also needs to demonstrate genuine business value, says CIO Vikram Nafde, CIO at Connecticut-based Webster Financial Corporation. “The cost of implementing and operating AI models can be quite high, so you need to be very careful in assessing the business worthiness of AI use cases,” he says. “This involves rigorous evaluation of potential benefits, risks, and costs associated with each AI initiative to ensure investments are prudent and aligned with our risk-return profile.”

Early definite results

At Eaton, several PoCs are already delivering results, and they’ve used AI to consolidate data across more than 70 ERP systems globally. Leveraging expertise from software developer Palantir Technologies, Redmond’s team developed a model that consolidated and cleansed the data from those systems, then analyzed it to produce insights — and fairly sophisticated recommendations — for decision makers.

As an instance, if manufacturing in a single substitute unit is short on eight-run steel rods wanted to assemble orders in meeting, and one other design of the substitute has 10-run rods readily accessible, the AI would possibly per chance indicate utilizing the longer rods and reducing them genuine down to blueprint the provision deadline. “A human reviews it to be determined it makes sense, and if it does, the AI contains that into the studying mannequin,” she says. The venture, which is mute in what Redmond calls the “value proposition stage,” has already generated definite results for the firm’s electrical substitute. “On-time offer has improved substantially,” she says.

Webster Monetary institution is following a similar system. “We’ve established an AI working group with representatives across technology, architecture, records, security, ethical, threat, and audit consisting of both technical practitioners and business users to assemble AI best practices and a governance framework,” says Nafde. The financial institution is also looking at AI to help streamline internal operations and gain efficiencies, including building customized items particularly tailored to the business’ needs. As an example, it’s experimenting with utilizing gen AI to automatically read financial statements from company possibilities when assessing loan applications.

“The teams here must read and process a large volume of financial records, and it’s almost never in a standard format,” he says. “Generative AI can read and extract the key data and summarize it for people.” To this point, he says, “We think it’s a strong use case. It can also be faster, more accurate, and make the teams more productive.”

Eli Lilly and Company is also at the forefront of adopting and integrating AI into the business. “We’ve found that AI can help in just about every way to streamline workloads and advance our research and development,” says EVP and CIDO Diogo Rau.

Currently, gen AI is helping with novel drug discovery by creating never-before-seen molecules and examining their potential in the development of new medicines. It compresses years of work into months — sometimes even days, Rau says. Lilly also developed an AI tool to monitor and clarify patient data from medical devices, and enhance the safety and effectiveness of medicines using a proprietary “sensor cloud.” Plus, it uses gen AI to automate the creation of initial versions of systems to generate documentation supporting clinical trials, and to generate materials for regulatory submissions.

Production is another design that benefits from AI. “At Lilly sites, we leverage refined algorithms and systems, automated guided vehicles, fully automated warehouses, robotics, and highly automated manufacturing equipment to increase and accelerate the production of our medicines,” Rau says.

Partnerships are key

To develop PoCs for AI initiatives, CIOs like Redmond at Eaton are turning to trusted partners for support. “It was essential because we don’t have a large pool of AI resources, and you need a model to start with,” she says. “It’s a real accelerator in the beginning.” However, she adds, it’s also important to bring your internal team up the learning curve to keep costs down as projects move into production.

Webster Financial is leveraging hyperscalers such as Microsoft and AWS whenever possible, and equipping the bank’s own expert technologists to build what’s most important for its needs, all while minimizing reliance on consultants. “That way, we don’t have to depend on expensive contractors for both development and ongoing support,” says Nafde.

Have confidence, but take a look

A successful PoC doesn’t guarantee success: Stakeholders prefer to believe it. At Eaton, as an illustration, an AI-based sales forecasting tool has the functionality to enhance productivity dramatically. Currently it takes months and thousands of man-hours for all its finance and sales groups to learn about historical records, combine it with fresh sales forecasting records, and blueprint projections. Now, says Redmond, “The AI model can potentially assemble that for you.”

Eaton’s forecasting PoC venture, which ran in Q4 last year, has been no less than as honest, she provides, if not more so than the fresh scheme. “It’s definitely better than what we’re doing now, spending thousands of hours working on it,” she says. The question is whether or not people will be willing to trust the technology enough to give up doing the work themselves. “We’re not there yet on people being happy letting it go,” she explains. “We’re still in the ‘trust but verify’ piece.”

Yet another reason why trust in AI could create a trust issue — both inside and outside of IT — is the truth that the model is a black box when it comes to understanding exactly how the output was determined. “And for the first time, with generative AI, we’re working with technology that’s not deterministic; it’s not binary,” says Sanjay Srivastava, chief digital strategist at Genpact. “For instance, with generative AI, you might get an answer that says 94% of the time it’s correct, which means it needs some oversight or augmentation.”

“These tools are incredibly efficient and sometimes convincingly sinful,” says EY’s Diasio. Unfortunately, though, there’s a tendency for folks to go on autopilot. People prefer to work with the tools and review the output, not just casually, but thoroughly. “You have to make the time for that,” he says.

Srivastava says most initiatives keep a human in the loop to make final decisions, but follow-through is indispensable. “How do you go from data, to insights, to action in a real loop?” he asks. “That’s the number-one reason why people don’t achieve economic outcomes.”

Data prep issues, rather than…

In areas comparable to offer chain and analytics, having your total records in a invent readily accessible to an AI mannequin is a must have. “Recordsdata is the lynchpin to AI success,” says Nafde. “Originate along with your records system prior to your AI system, and align your AI system along with your substitute system.”

Diasio is of the same opinion. “Invent determined the records you’ve is discoverable by AI programs, which would possibly well per chance mean building an enriched catalog utilizing generative AI or utilizing it to make an ontology on high of structured records,” he says. “In plenty of instances, it’s a vital development in productiveness when utilizing AI to streamline these workloads. In some records migration whisper we’ve noticed a 40% amplify in numerous steps along the model and an amplify in creep.”

Lilly is already using AI-enabled tools to speed the ingestion and cleansing of the data used to train and fine-tune its pharma products, Rau says, and Genpact also uses AI to prepare its data for consumption by its AI products. “We have a ton of data and two-thirds of it is unstructured,” says Srivastava. “You can use generative AI to automatically create a semantic layer on top of your data. You need to know what data sits where, how it’s linked to something else, what the quality is, the lineage, and where else it’s being used.”

That work is complex and requires highly skilled expertise, which is why many enterprises bring in a partner to help with the work. However, AI can automate the creation of that semantic layer for you. It’s not perfect, but it can get you to 80%, Srivastava says.

On the other hand, Diasio says you don’t often need to prepare internal data to leverage AI. “For example, with generative AI and the pre-trained models available, creative initiatives such as product design, or summarization initiatives like contact center transcripts, can also work successfully out-of-the-box in the right contextual setting and with effective prompting,” he says. “It will help companies use the power of AI while they continue to curate their internal data and harvest their insights.”

Ensure suitability of AI capabilities before activating them

“CIOs should invest in new or enhance existing CRM, IoT, ITSM and business intelligence tools that include AI/ML,” says Jevin Jensen, research VP at IDC. “Time to value is dramatically reduced when you choose a solution from an existing off-the-shelf vendor that has added AI features to systems you’ve already implemented.” You may simply need to turn on the feature or add a plug-in. Be sure to determine whether you can opt out of having your data used to train the vendor’s models, he says.

While new AI capabilities in enterprise systems such as those offered by Salesforce and ServiceNow promise significant workflow productivity benefits, you shouldn’t simply turn them on without fully understanding how they fit with your workflows. “We recently had a deep-dive session with ServiceNow on the specific way to use predictive intelligence, digital chat, and other capabilities in alignment with our business systems,” Nafde says. As an example, the bank’s digital chat feature involves a couple dozen use cases. Some will likely be ready to use it right out of the box, some would require customization, and some may not be suitable for purpose. “Now we need to determine which capabilities will be valuable,” he says.

Eaton has already begun using some AI features in ServiceNow, with encouraging results so far. “It’s helping from a case management perspective, finding threads of defects we can improve, finding the root cause, and offering solutions that could reduce case counts,” Redmond says.

The conundrum with embedded AI in enterprise systems, though, is it may not provide a compelling solution as of late to your organization’s needs. In this case, CIOs, especially if they face competitive pressures, can also find themselves in a dilemma: “Should you wait for your line-of-business application vendors to incorporate AI and sacrifice time to market while you wait for the supplier to make it, or should you build an enterprise architecture system where you have your own customized implementation and infrastructure around it, but it’s costly and requires ongoing investment?” asks Srivastava. “Therein lies the challenge.”

Lilly is also leveraging AIOps capabilities in its IT operations. AI-enabled tools include an incident detection and response system that quickly detects anomalies, predicts potential issues before they escalate, determines the root cause of failures, and assesses the business impact of technical problems. “For example, if the order processing system experiences delays, AIOps can quantify the impact on revenue and customer satisfaction,” says Rau. This allows the team to prioritize and get to the bottom of the most severe issue sooner.

What to assemble – and no longer to assemble

Whereas Webster Financial institution is still in the early phases of its AI journey, Nafde has learned a few things along the way so far: Get your data in order. Align your AI system with your business system. Put the ethical KPIs in place before you begin. Then start small, demonstrate proof of value, scale gradually, and educate and communicate with your stakeholders every step of the way, he says.

Equally crucial is to associate to fetch off the ground, nonetheless make out your team with the tools and ride to assemble and protect fresh AI capabilities. And don’t underestimate the prefer to make belief. “Finish ahead on your messaging,” he says. “Quiz skeptics, assemble metropolis halls, and have leaders step in.” As a minimum, there’s fairly a pair of alarm and general reluctance to web swap when fresh technology is launched. “The discipline here isn’t correct about AI,” he provides. “It’s a conventional swap management discipline.”

Be strategic and limit the number of initiatives you take on, advises Redmond. “Focus on a few things and go deep,” she says. Find trusted partners to help you start, and take advantage of AI capabilities your SaaS vendors have launched into their products — when they make sense. Don’t miss out on what’s already in your ecosystem, she adds.

“Custom issues,” says Rau. “Change is difficult, so CIOs prefer to lead a cultural shift by demonstrating the forward-thinking behaviors you’re looking for, and creating an environment that encourages learning and innovation around AI. Our biggest risk is that if our employees don’t adopt AI as fully as they could.”

Getting some wins under your belt, working with stakeholders using it until they’re comfortable with the new technology, is a solid confidence boost, says Redmond. “That will reduce the alarm factor,” she says.