The AI Marathon: Using Quick Wins to Fuel Long-Term Success
AI adoption will never progress uniformly across an enterprise. The early results from trailblazing companies deploying Generative AI are astonishing, from the generation of code, to the creation of new ideas, to the deflection of customer service calls, to AI-generated advice to employees that leads to new sales.
There are caveats to consider, however, as IT has the responsibility to ensure that these initiatives scale. The fear of Generative AI sprawl complicates technology planning as every department turns to IT with its specific list of favorite use cases. Considering the caution and hesitation I have seen in rapid, widespread AI adoption, I recommend creating a layered AI strategy that mixes quick and well defined tactical projects in parallel with strategic planning for the next 36 months. The goal is to show value and create excitement in AI adoption, while also reducing the friction in launching initiatives and minimizing technical debt. The result is sustainable, effective AI integration across the enterprise. As an architect, you’re uniquely positioned to lead this strategic balance, ensuring your organization’s AI adoption is both innovative and sustainable.
In the past six months, CEOs have tasked their technical leadership to quickly come up with an action plan to fulfill the CEO’s number one technology priority: AI adoption. More specifically, CEOs want the technology embedded in business processes as noted in this Deloitte CEO survey. There is a problem, however: Initial expectations were very high, and what originally seemed very achievable in a short time frame is now in question. Several factors have contributed to this new emphasis on a more measured pace of change, not the least of which is the fact that Generative AI projects require reliable and accessible data as their foundation.
The future of AI depends on reliable data. With relevant, accessible, and trusted data, AI is practically magical. The reality, however, is that 59% of customers don’t trust companies with their data. Studies done by BCG show clearly the impact of trust on business. In creating their Trust Index, they found that the most trusted companies generate 250% as much value as the average business in the same sector, and that these companies are rewarded with a price-to-earnings (P/E) multiple that is 47% higher. Getting the data required to supercharge business applications with AI into a form that is trusted and actionable is now an imperative.
Tactical vs. strategic AI projects
Insufficient access to actionable customer data is a key hurdle in engineering trust into technology and process design across a business. In an earlier blog post, I looked at the ways in which businesses are experimenting with the ideal balance of ownership and decision making among AI, Generative AI, data, analytics, the CIO, Trust officers, the CISO, and line of business owners (including Sales, Marketing, Service, and eCommerce). AI projects cannot run for long in a vacuum or in silos supported by disparate systems and strategies. Even if there are strong improvements to a few tactical performance metrics, the organization runs the real risk of destabilizing a comprehensive AI strategy meant to achieve four interconnected goals:
- Improve employee experience and performance
- Improve customer experience
- Reduce operational costs through greater efficiency
- Drive growth
One-off tactical AI projects can be positive: Many of them are yielding great results, such as helping salespeople sell better and enabling customer support agents to evolve into brand ambassadors who perform at a higher level. Numerous growth-oriented service organizations are introducing AI-powered next-best actions to augment agent decision-making and achieve a new level of fidelity in customer self-service through personalized and proactive responses. However, without strategic alignment, these successes can lead to fragmented efforts and missed opportunities for broader organizational impact.
Navigating the AI adoption curve
The current challenge is that, to use Gartner terminology, Generative AI for CRM has left the “Peak of Inflated Expectations” and entered the “Trough of Disillusionment”.
“Through 2025, at least 30% of GenAI projects will be abandoned after proof of concept (POC) due to poor quality, inadequate risk controls, escalating costs or unclear business value”.
- The ROI in AI (and how to find it)
A look at the upside is that 70% of Generative AI projects that emerge from a proof of concept (PoC) are succeeding. The insight is that with disciplined planning and testing, AI-powered CRM projects are enabling a complete reimagining of what a business application is.
Organizations that have continued to evolve their CRM strategies by seeing Generative AI as an additive set of technologies will leapfrog the competition. Too many others see the revolution in data and AI as a golden age for IT. It is not a golden age just for IT. Rather, Generative AI also heralds the golden age of CRM business applications.
The reality is that AI technologies are best harnessed to create better outcomes for employees and customers. Enterprise architects who foster tight collaboration between IT and key business units such as marketing, sales, and customer service are already in a better position to leverage advanced technology across the organization. This is because accurate, trusted, and available data is vital to every process that uses Generative AI.
Lowering the barriers to success
A Salesforce survey in 2024 shows that 88% of IT leaders feel they can’t meet AI demand safely. The survey reflects many other recent polls and surveys that list multiple challenges, including the top five:
- AI trust and safety
- Data trapped in legacy systems
- Lack of a data sharing model
- High cost of training Generative AI models
- Integration with business workflows.
Depending on industry and geography, the percentage of IT organizations that believe that their data is ready for a strategic AI program varies widely from as low as 4% to as high as 80%. The question is: Why are some struggling while others succeed? What is getting in the way?
- A missing skill set that blends AI expertise with business insights
- A lack of governance that negotiates the AI program goals with the needs of the lines of business
- A team that is able to prioritize the requests of the multiple constituencies who want Generative AI capabilities with the corporate business objectives
- The corollary of prioritization is introducing an ROI discipline that measures and tracks AI projects and measures new outcomes.
- Cost-benefit review that determines how to allocate funding such that the most critical initiatives advance unhindered.
There is a ready way to cross the chasm and break down the AI paralysis that some IT organizations are feeling. By following a few easy-to-onboard principles and creating a roadmap, there is no need for nervousness that AI is developing too quickly. In fact, it is developing too slowly! It is safe to dive into the deep end of the AI pool once a few lessons are absorbed.
A review of many types of Generative AI projects reveals that of those that make it through the initial PoC stage, the successful standouts have several factors in common. Dependable data is one such factor. Others are the rigors of a security review, oversight by the legal department, the ease with which outcomes can be measured, and the quality of the skills available to the organization.
In addition these factors, creating a comprehensive plan and roadmap is essential. Establishing a data governance framework and investing in training are also critical steps. Refer to the Generative AI Roadmap Template and AI Resource Gallery for help getting started.
Creating excitement for AI in the enterprise
Everyone needs encouragement that Generative AI is truly a differentiator. This is the best way to overcome the disillusionment that arises from tackling the complexities of AI implementation. The bigger vision is profound and needs amplification: Generations of business software were primarily systems of record requiring a steep learning curve. With the emergence of Generative AI, business applications come alive, advising, directing, and augmenting humans in new and exciting ways. Both employees and customers feel empowered with an advisor at their fingertips, often acting proactively on their behalf. For instance, digital agents — autonomous, intelligent agents — work in the background on behalf of customers, revolutionizing the human-to-digital experience in customer service, commerce, marketing, and sales. To outpace the competition, smart businesses will collaborate with partners focused on connecting the business with customers in innovative, AI-powered ways. There has never been a better time for the enterprise architect to lead the charge.
