Generative AI is an exponential technology—and integrating it into your organization will require going on an exponential journey as network effects build (of data parameters, users, and other key ingredients of large language models). Before jumping into generative AI with “what do we do next?” always take a moment to ask, “are we thinking about this right?"
The generative AI field is moving at stupid-fast speeds—at least, it seems like it is if you watch your LinkedIn feed. Like any exponential strategy, we need to prepare resources to scale even as we are still figuring out our exact destination. Sometimes the best way to close the resource gap for generative AI is to sort your needs into "build, buy, or partner" categories.
One way to decide between these options comes down to organizational DNA, or the core characteristics your company is good at and known for. The DNA of an insurance company, for example, is usually the ability to evaluate risk, whereas a news company might value its neutrality and fact-finding, a fashion company might value its trend-setting and exclusivity, and a hospital group its standards of care. Look to your DNA to help you make a decision between building, buying and partnering; straying too far from what your company has earned a reputation for is both unlikely to succeed and also risks damaging your credibility.
Building requires creating your own infrastructure and hiring and maintaining top talent, which can be very costly and time-consuming. If you are using generative AI to create something core to your value as a company, consider building it yourself to maintain a competitive advantage—provided you have the talent, time, and tech in place.
For example, Bloomberg has built BloombergGPT, a 50-billion-parameter system. The financial giant's DNA is gathering and analyzing data to produce information; it wouldn’t work to just patch ChatGPT into their systems and say they had an AI offering. Bloomberg spent massively to invest in a home-grown solution that promises to be very, very hard to compete with—earning them a first-to-market advantage for customers while creating further moats around their proprietary tech approach.
Building in-house requires a team of experts with knowledge in machine learning, natural language processing, and conversational design who can develop and implement the technology. This approach allows for total control over the product but also requires significant resources and time and is definitely an exponential initiative, with all the leadership and funding challenges that can entail.
Buying access to generative AI tech can be a good option for businesses with limited technical expertise or resources. If the generative AI resource you need is not close to your DNA or core value (say, an AI-powered customer service bot for a video game company), and is instead just a utility function you need to bring your company to parity with the market, consider buying off-the-shelf AI to start. While it risks leaving you dependent on vendors, it can get new digital offerings to market much more quickly, and it also often means that you can allocate more of your resources towards differentiated innovation (like in-game generative AI for characters, in the example of a video game company). When searching for a ‘buy’-able solution, carefully consider the original training data used and whether it will grow and adapt to your usage (like Writer’s customizable LLM approach for on-brand content) or stay generic (like ChatGPT). Organizations also should consider whether their IP will be comingled with others, a concern (and reality) which caused Samsung to quickly prohibit use of ChatGPT by their employees.
Partnering accelerates network effects quickly. For example, you might have a large customer base, while another company has financial infrastructure. Top tech companies partner for rapid growth, as Apple did when partnering with Goldman Sachs to create the Apple Credit Card. If the solution you need is part of your DNA but you don’t have the internal resources or want to leap to the top of the market quickly, it may make sense to deeply partner. In the AI space, we see software behemoth Microsoft—with its massive installed user base—bringing OpenAI’s technology into Office and Github seemingly overnight, while Google follows far behind (at least for now).
Closely partnering with (or acquiring) external companies has an advantage if they come pre-equipped with tools, personnel, experience, and potentially a customer base. They can provide insights into best practices for generative AI development and help identify potential pitfalls beforehand, reducing development time significantly. However, it's important to make sure there is compatibility between their existing processes and those of your company. Also, there are few mature generative AI companies to partner with at the moment, and the technology is evolving very rapidly; do due diligence before investing heavily in a partnership.
Whichever approach you choose should be informed by whether you are using generative AI to do a ‘faster, better or cheaper’ version of your current work or innovation which will express your DNA in a wholly new way. Table-stakes goals like increased efficiency and improved customer service experience are good fits for buying; building or partnering makes more sense when you’re creating net-new value from your core DNA. One way or the other, always ask, “are we thinking about this right?” before “what do we do next?”
This is one of a ten-part series on Questions to Answer in AI Strategy. MJ Petroni is a Digital Fluency and AI readiness speaker, author, and accelerator. You can read even more in the Digital Fluency Guide.