Industry-specific generative AI apps look set to be the game changer. Already, some legal and professional services firms are integrating Harvey, a generative AI platform for lawyers based on an enhanced version of GPT-4, the large language model used by ChatGPT. The international law firm Allen & Overy announced in February that 3,500 of its lawyers had asked Harvey around 40,000 queries for their day-to-day client work. In March, PwC announced a partnership with Harvey, adding that it would work with the startup to help clients use it and look to develop and train its own proprietary AI models.

Industry-specific apps bring the potential for collaboration across and within industries — and indeed, with their potential for pooling the proprietary knowledge of companies, could affect how both ecosystem co-operation and M&A are viewed. Industry-specific apps that segregate company-specific data while offering similar functionality to all customers are likely. We believe the explosion of interest in human-centric, generative AI in early 2023 will accelerate their arrival.

Building an AI strategy

Along with powerful new AI tools coming onto the market, big tech is prioritizing AI development, numerous startups are in the fray and expectations of AI are rising. The value of human-centered, easy-to-use AI is becoming clearer, fast. Bill Gates recently reflected on the tech produced by OpenAI — the company behind ChatGPT and GPT-4: “I knew I had just seen the most important advance in technology since the graphical user interface.”

Given all this, we believe it makes sense to revisit your company’s AI strategy, or draw one up if you don’t already have one. Here are some ideas that might usefully feed into an AI strategy, whether it’s for generative AI, human-centric AI, or the broader AI piece.

  • Organize an AI task force: The flourishing of human-centric, accessible AI merits the creation of a multidisciplinary AI task force, including strategy people, software engineers, data scientists, service designers, domain experts and legal, reporting to executive management. Even in companies that already have an AI team or experts embedded across departments, this may be a necessary step. Make sure AI is treated as top priority, not hidden as a subset of your data strategy.

  • Grow your strategic thinking: What use cases and associated AI capabilities are available to you and how do you want to use them? How could human-centric AI improve access to your company’s collective wisdom and experience? Which teams could it help to support and free up? Identify strategic pillars where AI could help, such as in decision-making by senior teams, customer experience, operations, or driving innovation. Assess which human capabilities it could be used to augment. Spend time scanning for downsides — will over-reliance on generative AI erode some important skills, or let in mistakes? More broadly, could AI mean your market expands, or bring risks to your business model?

  • Take a holistic, people-centered approach: Leveraging AI gainfully is about much more than building technical capabilities — partnering humans with AI can enhance people's capabilities. How could this change how the company is organized and how it runs? It’s crucial for business leaders, domain experts and technical teams to stay in sync and speak language they all understand.

  • Identify and capitalize on the relevant data: The AI task force should work to source data that can be used to train AI. For generative AI this might be from contracts, meeting notes, internal reports and company strategies, externally-provided data sets, customer interactions, images, video, audio, news feeds or social media. Consider whether any could be used to create externally-facing industry-specific apps (which may create new value streams). AI depends on accessing data so will be more effective if the underlying data is organized in an efficient and accessible platform. Siloed, poor quality data may go from being an annoyance to a critical business issue.

  • Strengthen governance and oversight: AI, particularly generative AI, can bring new legal risks that need C-level recognition, for instance around intellectual property, liability for providing incorrect information, and data privacy. Strong communication is needed between legal and tech teams, together with guardrails around AI’s users and use. Understand how mistakes made by AI would be identified and managed. The trade-off for fast, inexpensive content generation needs to be robust quality control.

  • Build an ethical approach, including on emissions: Research shows ethical and legal considerations are often disregarded in the design and deployment of AI systems. Understand any risks of bias in the AI you are using, together with uses that could be unethical. Training AI systems can require significant computing power, which means greater emissions. Consider tools to measure and monitor cloud usage, like Cloud Carbon Footprint.

  • Manage the expectations of your teams: Many knowledge workers are wondering what AI — and generative AI in particular — might mean for their day-to-day work and the shape of their careers. The CEO and AI task force should work with the comms team to explain its capabilities and dispel hype and science-fiction type fears. Feedback from stakeholders across the company should be encouraged.

Sparked in part by the launch of ChatGPT — and some other smart product development and marketing by big tech — early 2023 may turn out to be a tipping point for the use of AI in business. Human-centric AI looks set for an increasing role, whether in the form of AI assistants, generative AI, or easy-to-use bespoke applications. Industry-specific and proprietary generative AI apps look likely to make an impact.

AI for business is now developing at speed, and formulating an AI strategy will help businesses to navigate through the noise to the real business gains.

Thanks to the following for their contributions to this article: George Earle, Dave Elliman, Emily Gorcenski and David Johnston.