As investors and corporate boards assess the situation nearly two years into the AI revolution, it’s becoming clear that playtime is over. For companies, it’s time for business GPT at scale.

At this point, fintechs like Broadridge are spending countless hours helping clients unlock opportunities to transform their businesses with AI. One of the most important steps in this effort is convincing companies to move beyond consumer-focused ChatGPT applications to more sophisticated GPT solutions. But before companies make that major investment in an enterprise AI platform, they need a strategy that lays out exactly how they will integrate AI into their workflows and functions and scale AI across the business.

AI-ifying Productivity and Products

While every strategy will be different, based on the company, its business and its legacy technology stack, any comprehensive plan for scaling AI will have to address two fundamental items: internal integration and product development. Companies should start out by asking how they will apply AI to internal workflows to enhance efficiency and productivity, and how they will employ AI in products to meet client needs and drive revenue growth. In the fintech space we add a third question: How will we use AI to lower risk for our clients, many of which are highly regulated financial service providers?

Companies just starting out on their AI journeys should take these questions to an assembly of experts from senior management, IT and product development teams for a no-holds-barred “AI-athon” in which participants brainstorm ways to leverage these emerging capabilities. The ultimate result of this session should be several potential real-life use cases that can form the foundation for the company’s initial push into enterprise AI.

Although internal efficiency and product development are distinct initiatives with separate objectives, they are intrinsically linked when it comes to AI. Both internal and product applications of AI will, in most cases, be powered by a single business GPT platform. If done correctly, the two efforts will feed each other: Internal integration of AI tools will spawn experimentation and creativity that can ultimately lead to the creation of new products, and capabilities that emerge from the product development process can be integrated to create efficiencies and other benefits internally.

Required for Scale: Technology and Training

This year, Broadridge introduced an enterprise AI platform that powers both product development and internal implementation. The platform, which included a protected “sandbox” where employees could safely experiment with AI, was designed to promote self-service functionality by business groups across both objectives.

However, early in the planning process we realized that providing technology was not enough to drive adoption and scale. Within companies—and among consumers—AI engagement seems to follow a regular pattern. A relatively small group of early adopters eagerly starts using and experimenting with the technology. This experimentation phase is quite valuable, and can quickly reveal challenges and opportunities. For example, at Broadridge we estimate that experimentation by professionals within our individual business units (a group we call “tinkerers”) has already resulted in the creation of more than 170 AI chat assistants that are currently in use across the organization. We believe Broadridge employees who are active users of AI save an average of three hours per week, or roughly 1,500 hours per week companywide. Early experimentation within business units also helped identify opportunities for new AI-enabled products like BondGPT for bond market participants, OpsGPT for financial services operations, and Global Demand Model for asset managers.

However, adoption of the technology does not necessarily spread across the organization on its own. Rather, companies must deploy proactive strategies to move employment beyond the core group of power users to the broader workforce. To drive usage and help embed AI throughout the organization, companies should employ a combination of smart training strategies and the steady rollout of new AI use cases. AI integration should be added to the objective-setting process for all employees, with the goal of making it as mainstream as revenue, profit and DEI objectives.

The good news is that it should get cheaper and easier to roll out new use cases as time goes by. The marginal costs to build new AI applications decrease as business units and tech teams are able to reuse patterns and tools. For that reason, it’s essential that companies work with their fintech partners to ensure that the business AI platform includes foundational capabilities that enable reuse and standardized controls that make scaling AI seamless and safe.

Safety First

Anything adopted at scale needs rules of the road. Rolling AI out at scale with reckless abandon is dangerous. Safety must be the primary consideration, and the planning process for scaling AI should start with risk management. In addition to technical tools and guardrails to protect data and prevent mistakes and misuse, companies must build out internal governance mechanisms designed to identify and mitigate AI-related risks. Just as AI will be embedded in every part of the business, AI risk management must permeate every level of the organization, including the board of directors, the CEO and risk committee, and all aspects of corporate and enterprise risk management.

To recap, these are the elements companies need to scale AI: (1) a comprehensive strategy to embed AI in both internal operations and product development, (2) a powerful and well-designed business GPT platform equipped with capabilities to enable the reuse of existing tools and facilitate the creation of new applications, (3) training programs and other initiatives to promote the use of AI, and (4) robust risk management structures from the boardroom down to individual business units.

With these pieces in place, companies will be prepared to move beyond the initial phase of individual trial and error with ChatGPT and enter the new era of AI-driven business, safely, quickly and at scale.

Read the full article here

Share.
Exit mobile version