HP’s ongoing AI transformation has brought the company sizeable productivity gains, but the maker of computers and laptops sees more room to expand its use of AI tools across its workforce.
The company visited New York City this week to talk up its plans for HP IQ, a forthcoming AI orchestrator built into devices, set for release later this year. After the presentations, Prakash Arunkundrum, HP’s chief strategy and transformation officer, sat down with InformationWeek to discuss what AI means for the company internally as it seeks to develop AI-driven devices for others.
“Internally at HP, we’ve seen about 20% improvement in productivity wherever we have deployed some of these tools,” he said, while noting the breadth and complexity of his organization. Though HP now operates internationally and was born out of the old Hewlett-Packard that got its start more than 87 years ago, Arunkundrum called it an “OG” startup from Silicon Valley — agile enough to scale AI across the enterprise. (The original company split in 2015 into HP and Hewlett Packard Enterprise.)
HP’s internal transformation includes leveraging agentic workflows, as well as bots at employee workstations for customer support, with resources connected to the cloud and locally, Arunkundrum said. Even for a tech company, getting HP’s more than 50,000 employees to actually use the latest AI models has taken planning and effort.
“About half of them have some kind of agent tool already deployed. And obviously, the goal is to get to everyone — 100% of them,” he said.
Beyond internal productivity gains from AI, Arunkundrum said AI is also shaping customer satisfaction and potentially enabling new business models. That includes having customer information ready when a support call begins. “With a voice or chat interface, [when] now you call me, I’m not asking you 20 questions before I get to why your computer is not working,” he said.
As HP develops its AI architecture to create agent models, it has worked with players on the Silicon scene, including Sierra AI, to ensure they actually build automations, Arunkundrum said.
HP’s engineering teams are eager to expand their use of AI as they develop new resources for customers, he said. “They actually want more. So we don’t have an adoption problem [with them]. We have a demand problem.” That includes interest in using the latest models from OpenAI and Anthropic.
“We signed up for an exploratory partnership with OpenAI, to deploy some of the frontier models,” he said.
Other employees use AI resources to support workflows, while general knowledge workers may be more tentative about how AI can improve their work. Arunkundrum described the more tentative group as a small minority.
The escalation of AI across markets and industries has been a lot to digest, and moving fast is not enough, Arunkundrum noted. “In a space of 18 months, we’re talking a massive shift in how we’ve understood these things. So, speed is a sort of Holy Grail for us to think about experimentation, being able to try new things and pivot if it doesn’t work,” he said.
With that notion of being nimble, a device maker such as HP needs to figure out how to stake out its territory in the AI space, which many users access remotely. “You can’t interact with AI if you don’t have a physical layer. We are the physical layer,” he said.
