It’s widely understood among business leaders today that AI is becoming table stakes. For it to make a difference, there needs to be a strategy for how it will provide an advantage, and then those plans need to be put in place. Cisco Chief Strategy Officer Ammar Maraqa told me the strategy needed for an AI transformation isn’t all that different from adding any new technology or process to an enterprise—but the speed of AI advancement and the acceleration it can provide businesses make this transformation all the more critical. There’s more from my conversation with him later in this newsletter. We’re taking a summer break, and Forbes CIO will not publish next week, July 16. We’ll be back in your inbox on Thursday, July 23.Until next time. This is the published version of Forbes’ CIO newsletter, which offers the latest news for chief innovation officers and other technology-focused leaders. Click here to get it delivered to your inbox every Thursday.Artificial IntelligencePhoto Illustration by Timon Schneider/SOPA Images/LightRocket via Getty ImagesOn Wednesday, SpaceXAI—the rebrand of Elon Musk’s AI company—released Grok 4.5, its most advanced model. This version of Grok was designed for coding, agentic tasks and knowledge work, and the company says it can also handle financial and legal items. It was trained on datasets dealing with code, science, engineering and math, a company blog states. SpaceXAI CEO Elon Musk said on social media that Grok 4.5 is “Opus class,” referring to Anthropic’s frontier model, but faster, more efficient on tokens and less expensive. Grok 4.5 was jointly trained and released with SpaceXAI-owned Cursor.But Grok 4.5 isn’t the only new frontier model that’s being released this week. OpenAI’s newest models—GPT-5.6 Sol, and its less involved companion Terra and Luna models—will see wide public release today, OpenAI CEO Sam Altman posted on X. This model was first announced by OpenAI late last month, and the company complied with a request from the Trump administration to initially limit its release for security reasons. In a blog post about GPT-5.6 Sol, OpenAI calls it “our strongest model yet,” with advanced capabilities in coding, biology and cybersecurity. The model has two new reasoning levels: max, which allows more time for deep research and reasoning; and ultra, which uses subagents to complete complex work. The June blog post about GPT-5.6 Sol acknowledges the Trump administration’s current form of AI regulation—an executive order requiring AI companies to voluntarily participate in a 30-day government review process before public release of new models to assess “advanced cyber capabilities”—but OpenAI makes it clear that isn’t ideal. This kind of regulatory stop keeps the “best tools” from being available to users who need them, the post says.“We are taking this short-term step because we believe it is the strongest path to broader availability in the coming weeks, while we work with the Administration to develop the cyber Executive Order framework and a repeatable process for future model release,” the post said.OpenAI also launched an upgraded conversational model this week, called GPT-Live. This new model was designed to provide more intelligent, human-style conversation—complete with pauses and naturally flowing interactions. Notable NewsAfter several months of using AI agents, tech execs place a high level of trust in their more straightforward capabilities—including code writing and automated report generation—but remain leery around more complex tasks. Forbes senior contributor Janakiram MSV runs down the high points of the survey, published in MIT Technology Review Insights and conducted in partnership with Microsoft, which asked tech execs to rank how much they trusted agentic AI to perform a variety of tasks. On a scale from 1 to 100, business report generation and code writing received a trust score of 83.5 and 82.5, respectively, but service mesh configuration and troubleshooting scored 37.5, disaster recovery testing got a 43, and database migration planning earned 44.5. Almost half of respondents said their top concern with agents is accountability for their decisions, and close to the same number said theirs is the potential for inaccurate results and hallucinations. Three out of five respondents said they keep humans in the loop to address concerns.Bits + BytesHow To Develop The Right Strategy To Get Ahead With AICisco Chief Strategy Officer Ammar Maraqa.CiscoWith a variety of AI platforms widely available, just bringing one to your enterprise doesn’t automatically give you an advantage. Getting ahead with AI comes down to the best strategy to make it work at your specific business: Deploying it in areas where it can make a difference, getting wide adoption among key employees and truly understanding the long-term value proposition it brings to your company.I spoke with Cisco Chief Strategy Officer Ammar Maraqa about how to put this strategy together and make it work. This conversation has been edited for length, clarity and continuity. How can companies figure out the best way to achieve faster AI adoption?Maraqa: There’s got to be strategic intent to adopt these technologies. There’s a lot of top- down transformation required, especially for larger companies where you really need to have a voice across the company driving transformations. The second thing is—and this is where we see customers—you have to, at a base level, start investing in the infrastructure required to deploy this technology. Just as we were getting comfortable with chatbots, you’re seeing the agentic wave. That takes the problems we talked about and really magnifies them. Modernizing the infrastructure—it’s both the actual technology, connectivity, investing in security; and also getting the base-level infrastructure, data access, controls—all of these things need to be in place before deploying chatbot-like technology. It is incredibly more complicated when you think about agentic. But that’s almost a no-regrets move to most companies. Modernizing infrastructure is no longer a nice-to-have; it’s very necessary. The third piece is how you actually deal with the human element around enablement and training, really understanding how to augment workflows and deploy use cases where you do the work that says, this is the highest ROI value that I can go after. It’s a combination of how easy it is to deploy and how much productivity I can get out of it. When I talk to my peers, a lot of them are going through exercises of saying, ‘What’s unique to me? Where can I deploy the most value?’ Where I’ve seen it deployed, it’s remarkable to be able to augment humans and take a bunch of lower-level work off the plate so people can have more opportunity for higher-level work. The other piece I’ve been talking to customers and peers about is, instead of thinking about AI taking the work you’re doing today faster or more productive, you can deploy AI to do the work you never get to. Finding the right ways for the right function, the right sort of messaging [and] talking points, but also real value would go a long way.How do you get people to the point where they can work with AI, drive strategy, and not be afraid AI is going to take their job?You have to do a lot of enablement and messaging from a leadership perspective. The other thing we’ve seen be very successful is make it a positive experience for employees. We do hackathons where we bring in a non-technical function—our finance function or our legal function. Create these workshops, make it fun, and then brainstorm ideas of how you can use this technology in the day-to-day lives of the folks. Every one of those hackathons have been incredibly positive for the employees because it’s very specific. You end up coming out with a set of ideas or tools that you can use, and you get a lot of feedback about what works and what doesn’t.We also do a lot of measurement: adoption, engagement, values as much as we can. For sales, a good metric is how much more time are you spending in front of customers, because we know that’s a big pain point for. Pick metrics that really matter to the individual doing their role, and be able to showcase how a particular tool gets you there.The more information you share, the better. When you step back and think about the amount of information people are getting about displacement of jobs or the end of the world, it’s not easy for them to translate how it works with their business. That gap can be bridged by having folks actually experiment, put their hands on it.What advice would you give CIOs for getting the right strategy, putting it on the right track and moving forward?Don’t think about your strategy as only building. You have to think holistically about all the levers you have to execute on your strategy and accelerate your time to market. Stay true to the initial intent of whatever strategy you end up putting together. You have to basically enumerate the assumptions you’re putting down while developing your strategy. Given how fast the market is moving, you have to constantly reevaluate these assumptions. If you end up changing your strategy every three to six months, I would argue you probably didn’t have a strong strategy to begin with.Spend the time to really understand your base assumptions and test them in a way that is very credible and almost dispassionate. Based on that, make changes. Don’t react without really understanding what specific assumption is changing, and how you change your strategy in order to react.Comings + GoingsRide-hailing company Lyft appointed Senthil Padmanabhan as chief technology officer, effective July 20. Padmanabhan will join the company from eBay, where he has worked as vice president of engineering.Restaurant franchise chain International Dairy Queen hired Phil Crawford as executive vice president and chief technology officer, effective July 7. Crawford steps into the role from Adyen N.V., where he led the global food, beverage, and hospitality verticals. He succeeds Kevin Baartman, who is retiring.Dessert company Crumbl promoted Jacob Moncur to chief technology officer. Moncur joined the company in 2020 as vice president of engineering, and he will succeed Bryce Redd.Strategies + AdviceYou want your employees to use AI, but you want to make sure they’re not getting too friendly with it. It’s not uncommon for workers—especially those who say they feel lonely—to consider AI the same as a human co-worker and turn to it for emotional support. Here are some ways you can introduce some friction into the worker-AI relationship and steer your employees toward more human contact.Yes, AI can do tasks previously handled by junior-level employees, but that doesn’t mean you don’t need them anymore. Now is a good time to bet on recent college graduates, whose personal use of AI, enthusiasm, perspective and youth could be a bigger benefit to your organization. QuizT3 Magazine just celebrated its 30th anniversary and honored what it called the biggest tech product in its history. What was it?A. The BlackBerryB. The iPodC. Windows VistaD. The Motorola StarTAC phoneSee if you got the answer right here.