An onslaught of early generative AI products has kept the tech media buzzing for the better part of the last year. Case in point is the Washington Post story "AI Hype Is Gripping Corporate America." It's undoubtedly true.
However, beyond the shiny allure of the improved productivity GenAI promises is a less exciting reality. Most businesses are not ready to embrace GenAI as a part of their daily workflows (or indeed don't know where to even begin), and there is significant work to be done before experimental projects turn profitable.
What's encouraging is that beneath the juxtaposition of excitement over the new tech and the reality of implementation lies budding innovation that technology pros — who are truly ready to maximize the potential of GenAI — can ride right to the top and deliver business value.
The Generative AI Gold Rush
GenAI startups, even in their most nascent stages, are flush with capital as investors jump on the trend, eager to get in on the ground floor. According to CB Insights data, between January and August 2023, $14.1 billion in equity was invested in GenAI.
What will turn these early-stage investments into game-changing, profitable solutions is a well-thought-out adoption plan, from use case ideation to deployment. But among all the noise, where do you start when so many other things take priority? To avoid the paralysis of too many choices, the following steps can help chart out your organization's GenAI path toward profit.
Start with 'Where'
Before making a significant investment in GenAI, survey where within your organization it does and doesn't make sense. Where can GenAI create efficiencies by making manual work easier and eliminating redundancies? What rote and repeatable (and in some cases boring) tasks managed by humans could be better performed by GenAI, so employees can focus on more fulfilling and higher-value projects? Can you build GenAI into your sales workflow to make cold calling more personal or use it to identify financial anomalies that indicate future success or failure? As GenAI use cases are being devised, it's equally important to assess the status of your data sources and systems, which are critical to train and power the large language models (LLMs) that are the underpinning of GenAI solutions.
Also, it's important to recognize that GenAI can become a cost center quicker than a profit center if not implemented correctly. Without thoughtful execution, GenAI is likely to face a similar fate as widespread cloud adoption more than a decade ago — often over budget and behind schedule. For example, teams that fail to survey their data first risk having to reconfigure their GenAI solutions down the road because they were not optimized to meet the end goal.
A final matter to consider before starting with GenAI is who will champion the cultural change necessary to make full GenAI adoption possible. Although many have heard of GenAI thanks to industry headlines, they don't fully comprehend how the technology works or its potential impact on their jobs or the company's business processes. Identify those key stakeholders and ensure they are aligned on the top-line goals so they can continuously help with transitions in your corporate culture.
When developing use cases, be cautious not to lean on GenAI as a crutch to support an under-resourced team. Though widespread misconceptions may suggest otherwise, the technology is designed to enable already hard-working teams to be more efficient, not replace team members.
Build or Buy?
Once you understand your organization's most pressing needs, the next step is to decide whether you have the core competency for developing and training AI models or if you need to bring in an outside partner.
Evaluate GenAI partners like you would any other vendor: Look for a proven track record of delivering value, seek out the partners that have done work you'd like to replicate, and go in with clear deliverables and a timeline established in step one. Front runners will emerge, but the best partner for your GenAI transformation will be one that's fit to address your specific needs and strategic goals.
When buying a GenAI solution, the same principles apply. Match up your options with your company's goals and circumstances, clearly understand how long it will take to integrate the solution into your existing workflow, and ensure that you have the internal resources that can bring a GenAI project over the finish line.
To either build or buy GenAI solutions and expertise, you need a robust data strategy. Without sufficient and organized data to power LLMs at scale, your GenAI applications will sputter out in the pilot phase. Harvard Business Review's Tom Davenport and Maryam Alavi explain the three major ways to incorporate proprietary content into an LLM: training an LLM from scratch, fine-tuning an existing LLM, and prompt-tuning an existing LLM. Ensuring that your data is buttoned up will increase the likelihood of your GenAI solutions becoming profitable.
Don't Launch Before You're Ready
Another potential GenAI pitfall that can hinder long-term success is scaling or launching a product before you're ready. Iteration and testing are key, especially as this technology continues to bloom, to ensure you are not wasting resources on a product that ultimately doesn't solve your problem.
Launch pilot programs within a few limited sectors or groups of your company before making the solution widely available or commercializing it. At Persistent, we soft-launched a GenAI-powered HR solution internally before we took it to market, and this trial phase gave us time to work out the kinks in a relatively low-risk environment before bringing it to the masses.
Don't Forget About the People
Your technology is ultimately only as strong as the employees who take advantage of it, build it into their workflows, and commit to understanding how to improve it over time. A solution that is not designed with your existing workforce in mind is bound to fail. To truly transform your company using GenAI, dedicate a portion of your resources to training your teams. This not only ensures that they feel confident embracing the new tool but that they are fully leveraging the solution's whole capabilities. It also helps employees realize that GenAI-powered tools can be incredibly helpful in their jobs and can open up opportunities for initiating new projects and learning new skills.
As GenAI is introduced to employees and users, company leaders must protect against any inherent or perceived biases in their GenAI models, and educate their teams on how to recognize them. Establishing a GenAI Center of Excellence (CoE) can help safeguard against potential bias issues and help create critical guardrails around how LLMs access and utilize company data. A CoE can also function as a clearinghouse on issues related to risk, security, compliance, and transparency so your company is demonstrating responsible AI practices. Include employees with diverse specialties and backgrounds within the CoE to create a holistic perspective on GenAI adoption, ethics, and learning.
Amid the GenAI gold rush, where headlines herald the promise of a technological revolution, it's crucial to pause and reflect on the key takeaways from this frenzy: Many businesses are unprepared for seamless GenAI integration. To unlock its potential, start with a well-considered adoption plan. Identify where GenAI can boost productivity and efficiency but be wary of execution pitfalls and ensure your data is up to the task.
Selecting the right champions to drive cultural change is crucial. GenAI's success lies not just in embracing the technology but in careful planning, thoughtful execution, and a commitment to empowering your teams for the long term.
Dr. Rajesh Gharpure is chief delivery officer at Persistent Systems.
