There is so much to learn and so many concerns to address.
Are you as confused as I am about artificial intelligence (AI)? The technology seemingly came out of nowhere (it didn’t really) in the fall of 2022 with the introduction of OpenAI’s ChatGPT. At the January 2023 Executive Connection Summit (ECS), it was the subject of a thought-provoking presentation. Fast-forward to ECS 2024, and multiple presentations addressed this steaming, hot topic.
Meanwhile, the AI evolution continues. HP spent much of its HP Amplify 2024 Partner Conference talking about AI, including introducing AI-enabled PCs and an AI Master Class for its partners to better help them understand how to match AI-enabled technology with their customers. The company is also looking to elevate its print customers’ printing experience by training AI to deliver perfect print. The company plans to launch this solution later this year. Called HP Print AI, HP has trained its print model to only provide optimal outputs, reducing wasted ink and paper.
Expect many more imitators and innovators to introduce AI-enabled products in the coming months and years as more businesses learn to leverage existing AI technology to improve business functions. As Patrick Gelsinger, CEO of Intel, observed during the HP Amplify Partner Conference, everything will be AI-enabled, and Intel is building AI into its products. Similarly, Cristiano Amon, CEO of Qualcomm, emphasized that use cases for AI are evolving quickly; however, we don’t know all the use cases because of how fast the technology is moving.
Great Minds Think Alike About AI
Great minds might not be thinking exactly alike, but they are thinking about it, and more of those thoughts were on display at the HP Amplify 2024 Partner Conference. In addition to Intel and Qualcomm, leaders from Microsoft, Google, NVIDIA, and Advanced Micro Devices (AMD) informed attendees about how their organizations are developing AI-enabled products.
Satya Nadella, CEO of Microsoft, discussed the changes AI is driving at Microsoft, including a new user interface. “We are creating a computer that understands us,” he revealed. “It has a new reasoning engine to make sense of what we digitize.”
He added that the value of new PCs, particularly those enabled with Microsoft Copilot, is to help people create better and faster. “This is a new age with expertise at your fingertips, information at your fingertips,” emphasized Nadella.
Google CEO Sundar Pichai discussed how AI can enhance creativity and productivity, predicting that future generations will expect a powerful collaborative experience for creation. For HP’s channel partners, he identified a big opportunity within their internal organizations—rethinking workflows by leveraging AI.
Alex Cho, president of personal systems for HP, Inc., highlighted the connection between AI and hybrid work and what it means to a technology company like HP. “Our mission is not about providing gear; it is enabling these two mega-trends to be synergistic—hybrid work, how people are working, and AI as a way to help enable that.”
Cho also maintained that security will be even more important because of AI, but it will be more than just security; it will be security and privacy. “And in that landscape, we have invested several years in our security platform, and we’re going to take it to the next level,” he said referring to the first platform that protects against quantum computer hacks.
Misconceptions about AI
Last year, former MPS advocate West McDonald started a new business, GoWest.ai, focused on artificial intelligence. In less than a year, he has become the office technology industry’s foremost AI proponent and has partnered with Keypoint Intelligence, assisting the company in its AI initiatives. One thing McDonald has discovered by talking to dealers about AI is that they have some significant misconceptions about the technology. One frequent comment he hears is that they are already using AI because some employees are using ChatGPT or Microsoft Copilot.
“The biggest misconception is that AI is only good for writing, and it is, honestly, the lowest ROI of what AI can do for them,” opined McDonald. “Most leadership teams I speak to for the first time about AI and business applications have no idea how it can solve workflow issues in every department and that the ‘writing’ element of LLMs [large language models] is only one application.”
He emphasized that AI can be used effectively by sales teams to automate CRM updates, create and manage tasks and follow ups, and summarize sales meetings for their managers to review on the fly. Service departments can benefit by creating custom GPTs to shorten troubleshooting time and marketing teams can use AI to get instant insights and sentiment analysis for web activity and surveys.
“Our AI workflow assessments are one way for them to understand how they can apply AI in their business,” said McDonald about his organization. “The workflow savings from applying AI to specific workflows are mind-boggling.”
Learning Curves
Dealers have many questions about AI, and McDonald has been fielding those questions frequently lately. Hands down, the top question is, “Can we sell AI and make money?”
“Dealers are good at selling and servicing ‘things,’” observed McDonald. “Those things could be printers, copiers, cybersecurity, managed IT, etc. So, they want to sell AI. AI models are not things they can sell today. Sure, some purpose-built offerings have reseller models, but honestly, dealers aren’t going to make a ton of money there.”
He noted that most LLM providers like OpenAI, Microsoft, and Google, all sell direct, so there is no channel opportunity. “There are ways for dealers to make money with AI services, supporting their customers’ AI journeys and getting paid to keep them productive, happy, and safe, but before they can do that, they have to eat their own dog food,” suggested McDonald. “Dealers need to first figure out how AI can save them money, give their people more time back for meaningful work, and automate things that are manual today. There are untold dollars and hours they can save if they take a strategic approach.”
He recommends dealers do this first because when they get good at it, they can do the same for their customers. “Everybody needs AI assistance from a business perspective, and they can build models around what they learn from doing it themselves first,” observed McDonald.
Another question he is often asked is, “How can we use it safely?”
“This is one dealers should be worried about,” cautioned McDonald. “The bad news? By ignoring AI, by doing nothing at all, the biggest risks to their data security and IP are happening right now. They may think their people aren’t using AI, but many are, and not safely. They are using free versions, which use any and all conversational data to further fine-tune and train [the AI developers’] AI models. Yes, this can be turned off, but it’s not intuitive, and people aren’t trained on how to do this, so it never gets done.”
McDonald explained that one of his first tasks when engaging with a dealer is to help them build an AI guidelines and policies document so that people are trained on what they are allowed, and not allowed, to do with AI. “The second thing we do is get them to standardize on a business-grade level of ChatGPT, Gemini, or Copilot. Why? Because the ChatGPT team, for example, never uses data for training.
The company also gets a private workspace with much more security, similar to what they experience when they use Office365 or other enterprise-level AI models. You pay a little more, but in the long run, the security and ability to standardize use in the office offers way more ROI and peace of mind.”
More savvy dealers or dealer team members often ask McDonald how they can trust AI if it hallucinates and makes things up. “Large Language models like ChatGPT, Claude, and Gemini have come a long way when it comes to hallucinations and making things up,” he observed. “But if people are using AI blindly, it can still happen, and it can still negatively impact corporate reputation. This problem can be 100% eliminated if users are trained properly, are using the right AI models, are using it primarily for internal use, and have good human oversight.”
McDonald has built custom GPTs for people that list every source from which it obtained its answers. A user can simply click the link to the webpage, validate, and move on. “Also, having a human fact check is nothing new; good organizations have always had an editorial team that does this, even for human writing,” noted McDonald. “Can you trust AI? Yes, if you are using it correctly and internally.”
Another question is can you create a chatbot to answer customer questions? “Yes,” said McDonald. “Do I recommend it right now for most dealers? Nope. Most dealers are sales organizations, and questions are answered in the sales process. There are exceptions to this, for training chatbots (internal only), service chatbots (internal only!). This answer will change over time as the mystery of hallucinations is solved, but for now, why risk it?”
AI in the Dealer Channel
Earlier this year, Tom McMahon, president of Milner, headquartered in Atlanta, Georgia, attended the Executive Connection Summit. Since then, he tells me, AI has ruined a few of his weekends. “I’m spending all this time on YouTube trying to learn what’s going on and how we can best improve and how you develop these scripts and use it, and that’s just evolving like crazy,” he said. “We’re probably ahead of a lot and behind a bunch, and I wouldn’t say we’re experts.”
Milner’s IT business uses AI to remediate service tickets. It’s also being used for scheduling and dispatch. The team also uses AI note-taking apps, which then summarize those notes. “We’re looking at it now to help us with some of our human resources issues, onboarding paperwork and order processing,” said McMahon.
Asked where office technology dealers will see the most dramatic implementations of artificial intelligence in their businesses and/or from their vendors during the next 18 months, McDonald took a conservative stance. “I’m not 100% convinced (I love to be proved wrong) that vendors applying AI to their printers and MFDs is going to have any dramatic impact. For dealers, the biggest bang for their buck in the next 18 months is to use AI to augment their tedious and challenging workflows, department by department.”
He added, “They’ll make their people happier as they can do more meaningful work with the time they get back, customers will be happier because dealers’ service teams and others will be able to serve them much more efficiently, and investors will be happy because of the dramatic impact workflow-directed AI application can have on their bottom line. More work, less time, less money. That’s the magic of AI for the foreseeable future.”
The secret to the future of AI, or at least the short-term future, is in the workflow. “Workflow is going to be impacted in even more positive ways once we progress from LLMs to LAMs [Large Action Models],” predicted McDonald. “The AI will learn how to connect workflows and applications without complex and expensive API integrations and software development. The coming LAM models will shake up how we get work done even more than what we’ve been able to do with AI already. It is going to be epic.”
AI and Data Analytics
One area that could see a boom from AI is data collection and predictive maintenance. Ed McLaughlin, co-founder of Predictive InSight and former president of Sharp, has been involved in the office technology industry for decades. He was at ECS and sat through all of the presentations on AI. Recently, I asked him for his thoughts on AI. He referenced the emergence of ChatGPT and the content creation side of AI, which has been receiving a lot of press. However, he doesn’t believe content creation is where the importance of AI lies, even though he believes that benefit cannot be dismissed.
“I’m looking at AI very differently,” said McLaughlin. “I don’t know enough yet to be effective with it, but I want to see the analytical stuff, the Palantir type of analytics that solves business problems and improves how things are done and analyze all the different pieces. That’s when you get the predictive maintenance…and deliver things more effectively and answer questions about the accounts more effectively.”
McLaughlin acknowledged that the office technology industry isn’t working on those types of predictive analytics yet. “The data is there to do it, but we haven’t picked the right partner. The big data people like Palantir are doing it to solve massive military and enterprise-wide company problems. The shortfall is that the hardware to solve those problems is limited and very expensive for most people, particularly on our side of the business. But that’s where the real opportunity lies.”
Editor’s note: Palantir Technologies is a software company that specializes in software platforms for big data analytics and is currently including AI technology in its solutions.
