Listen in to any corporate conference call, whether the company is in retail or telecoms or finance, and there’s a good chance you will hear: “We need to be a tech company.” But what does that mean?

For a start, here is what it does not mean. Just as money cannot buy happiness, it cannot turn a company into a tech-savvy star. Investment is necessary but far from sufficient: Not all tech initiatives pay off. Nor does being an early adopter mean much; indeed, that can become a money pit if the investments are in tech alone.

Here is what becoming a true tech company does mean: deploying tech in a way that boosts profits and performance. For that to happen, organizations often need to build their own technology, not just make good choices about what to buy. In domains where competitive advantage is at stake, the greatest benefit comes from building unique capabilities. This can mean creating something from scratch or assembling something new from components that work specifically for the company. That is the conclusion of nearly a decade of McKinsey research into hundreds of companies that are true digital leaders. Here is one telling statistic: 70% of the digital leaders created their own software to drive operations.

What else matters in terms of truly becoming a tech company? Four principles stand out.

The 4 ingredients to a tech company

First, and most important, business and technology teams need to work well together, with technology embedded in company operations and culture. There is a big difference between a modern digital/AI culture versus the 20th century information technology (IT) culture. That difference can be summed up in one word: “requirements.”

In an IT culture, the business team hands off requirements to the technology team; both are important but operate separately, and the technologists are seen as support. At successful digital/AI companies, on the other hand, business and technology teams are seen as equal in importance. Rather than passing requirements documents back and forth for one-off projects, they own problems together and on an ongoing basis. In our 2023 book, Rewired: The McKinsey guide to outcompeting in the age of digital and AI, my co-authors and I found that of the 50 banks we benchmarked, only 25% had turned their digital investments into significant value. What set these banks apart was not how much they spent or what technology architecture they chose, but how well business and technology teams worked together.

Second, speed counts. Companies that are on the way to becoming tech companies iterate faster and faster. They complete the product release and update cycle in weeks, rather than months. They work in agile, short-sprint cycles. To make this happen, they don’t just hire cool kids from Silicon Valley and change the dress code. Instead, they reskill and upskill their workforces to create lasting impact, creating technology career ladders for their digital talent, so that technologists can learn from other technologists. Critically, they are less dependent on outsourcing. Their technology executives are doers, not vendor managers.

Third, tech companies adopt and scale well. They dig into how a business domain can be transformed. They think hard about what incentives can encourage business teams to co-own tech initiatives. Companies can build amazing technology teams in the most traditional industries and some of the least obvious locations if they approach it with the right mindset.

Finally, leadership is critical and starts from the top. CEOs need to understand not only how technology could reinvent their business, but how to change their company to harness technology. They resist the “everything all at once” temptation to focus on the most important domains. Determining whether their organization is functioning as a true tech company requires leaders to ask difficult questions. Which business domains are best positioned for technological transformation? How can the company attract the talent it needs? Is the tech talent road map as detailed as the road map for scaling technology? How many senior leaders would self-identify as tech capable? How can the company use its data to create competitive advantage?

Gen AI adds new pressure

With the advent of generative artificial intelligence (gen AI), becoming a tech company is even more urgent—and complicated. What makes gen AI so hard is that it can look so easy. Almost anyone can fire up an impressive looking pilot, but turning that into a robust, repeatable, safe, and scaled business impact is much more difficult.

Most companies understand that. More than half say they plan to build their gen AI capabilities internally, through upskilling, reskilling, and redeploying talent. And spending on gen AI rose sevenfold in 2023, even though tech investment as a whole fell. Nevertheless, in a survey of almost 900 companies done earlier this year, fewer than 5% said that gen AI was contributing more than 10% of their organizations’ EBIT.

While gen AI is new and exciting, the lesson from earlier transformations remains relevant: technology adoption for its own sake doesn’t create value. Competitive advantage comes from building unique organizational capabilities that enable companies to innovate, deploy, and improve technological solutions at scale. Simply adopting off-the-shelf gen AI tools will likely not be enough: after all, the competition can do the same.

In short, true tech leaders build their own expertise, and in doing so, ensure that their tech investments show up on the bottom line.

This commentary is from McKinsey & Company, a Fortune Global Forum Knowledge Partner. Rodney Zemmel is a senior partner.