Honeywell, a company that builds automation technology for buildings and aircraft, has a plan to create more than $100 million in value from using generative AI, a leading company executive said.
Sheila Jordan, the company’s chief digital technology officer, who helps lead the company’s AI activities, revealed that number during a conversation moderated by VentureBeat at an event in Atlanta last week focused on how generative AI is impacting security.
In follow-up, clarifying remarks, Jordan told VentureBeat that generative AI is already delivering “tens of millions of dollars” of net value, meaning the benefit created minus the costs. She said this is the net benefit already impacting the company on an annual basis, and that “north of $100 million” is in “line of sight.”
Jordan said the actual number will go up as the company’s 24 projects around generative AI yield further results, and the company adds even more gen AI projects to the list that the company hasn’t even thought about yet. “It’s exciting,” she said. “I do believe it’s going to deliver more incremental value. We’re just getting started.”
The comments are significant because outside of the big generative AI technology providers like OpenAI and Microsoft, few Fortune 500 end-user companies are on record saying they’ve created so much net value. Many end-user companies are highly regulated, like banks and hospitals, and have been slow to deploy the technology to end-users.
Jordan did not divulge which specific gen AI application is yielding the most value, but said the company has a range of generative projects that fall into five buckets (see the slide below for an illustration), and that value is being realized across all of them. She said GitHub and large language models (LLMs) for operations, in particular, have shown early promise:
Microsoft 365: Utilizing Microsoft’s Copilot alongside their core productivity suite.
Github: She said 3,000 Honeywell engineers are leveraging Github’s code generation functionality.
Large Language Models (LLMs) for operations: Honeywell is using OpenAI models on Azure to improve experiences in contact centers, for example creating technical publications for contact center agents; extracting data from legal contracts; and building a “sales assist” to help salespeople become more knowledgeable.
External applications: Honeywell is leveraging the latest generative AI features provided by third-party applications. This includes Moveworks, an application the company uses to help employees answer any of their questions, such as “How many PTO days do I have left?” Jordan said that question might sound simple, but answering requires knowing your employee ID, and searching systems to know how many days you’re allowed. Honeywell carries the same expectation for more generative AI features from applications provided by Adobe and Siemens, Jordan said.
Honeywell’s products & services: This is probably the most important area, Jordan said, since this is where Honeywell can inject differentiated generative AI features into its own products, like Forge. Honeywell Forge connects data across buildings to improve efficiency, safety, and security, and Honeywell is building generative AI into Forge.
Jordan’s comments (see her remarks in video below) echo the remarks made by most of the enterprise executives participating in our AI Impact Tour, a series of events we’ve been hosting across the country to focus on how AI is being put to work. There’s a consistent pattern. Like the other leaders we’ve heard from on the tour – from Wells Fargo, Citi, NewYork-Presbyterian Hospital, State Street, and Biogen – Jordan said she leans more positive when it comes to generative AI, versus neutral or negative because she sees the technology as a game-changer much like mobile phones and internet search. “The more I got into generative AI, the more I thought this really is another one of those disruptive technologies. It’s fundamentally going to change everything.”
One side effect of the excitement around generative AI is an “insatiable demand” from internal employees, Jordan said. (This internal demand is something that was also referred to by our previous tour speakers, for example, those from Citi and State Street).
Also like those other leaders, Honeywell’s Jordan said it’s important for organizations to create an overarching strategy around generative AI, to prioritize how to generate value, and to put controls around it, to make sure it is well governed, secure, and ensures privacy of data.
In its case, Honeywell has created a Generative AI Council, made up of representatives of the company’s functional and business departments. Each function and department has a plan for generative AI, and that translates into 24 active programs that have either been deployed or will be deployed over the next couple of months, Jordan said. Jordan is personally tracking the P&L of the projects, and how they are controlled, and works closely with a colleague who is running the Council. Moreover, the CEO holds an all-day staff meeting monthly, and generative AI is an agenda topic every month.
According to Jordan, in organizational change situations, you typically have about 40 percent of employees who are “ambassadors” for change, getting behind new trends, and proposing ideas. Another 40% of employees take a “middle” stance, and 20% are “naysayers.” She said it’s important to have forums including all three groups, and where the naysayers’ concerns are heard, which allows the group to move forward: “This really builds the flywheel effect,” she said.
Will generative AI replace jobs at Honeywell? Jordan also answered this the same way many other leaders on our Impact Tour have: “My hypothesis is that it’s going to replace the part of our jobs that is tedious, repetitive and that that we just don’t want to spend our time doing,” she said. That means workforces need to be more critical thinking and focus on decision making. Leaders need to nurture an environment that allows them to “push down the data, push down decision-making” to others.
Jordan also advocated centralizing crucial decision-making around technology architecture and data. She said a single internal organization should have end-to-end responsibility for controlling the company’s core generative AI projects, and shut down “shadow IT” operations, and ensure compliance with regulation. A sound data strategy is also key. Honeywell uses Snowflake for its data warehouse and for running everything on top of it. The company is now spending time on how to architect “what to do with the output of generative AI” in the organization. The company is fine with employees experimenting with public generative AI tools like ChatGPT outside of work, and generative AI features offered by third-party apps, she said.
She said Honeywell is using a generative AI-driven copilot from Microsoft to bolster IT security. Even before the generative AI breakthrough in late 2022, Honeywell had been using more traditional AI and machine learning techniques to create “actionable insights” around possible security attacks or other security incidents, she said. But she said generative AI is even more powerful, and the company is using Microsoft copilot to help summarize and streamline information, allowing security teams to be more efficient and to know where to spend their time. She said generative AI offered a way for companies like Honeywell to consolidate and streamline their security applications.
As for risks around generative AI, Jordan said she was “super worried” about the increasing sophistication of deep fake technology around voice. “There’s not a lot of technology that can validate voice in the enterprise IT environment,” she said, noting that while that technology is getting better, voice impersonation “is a real threat and it’s something we don’t have 100 percent covered yet.” She said technology is good at validating and securing text and encrypting data, but worries that generative AI can replicate her voice. Another speaker at the event, Kelly Bissell, corporate vice president and deputy CISO, Microsoft, agreed that deep faking is the biggest risk around generative AI, but included video along with voice. He said the industry has to do better at deep fake detection.
Overall, I thought Jordan’s comments about Honeywell’s experience provided insight for technical decision-makers around gen AI implementation. It reflected and confirmed many of the best practices we’ve been hearing on the AI Impact Tour: A structured approach that focuses on value creation, but also on effective guardrails and change management.
Consider applying to attend our next AI Impact Tour on May 8 in the San Francisco Bay Area, where we are again featuring how to put real GenAI applications to work at scaled companies.
(Copyright:VentureBeat Honeywell exec reveals plan to deliver $100 million in value with generative AI: "Just getting started" | VentureBeat)
Chinese
Japanese
Korean
