Balancing your dreaming and doing is key to enterprise GenAI success

When it comes to GenAI, enterprises must stop dreaming and start doing. Our recent Market Impact Report found that approximately two-thirds of organizations remain GenAI ‘dreamers.’ However, even after enterprises begin implementing GenAI to become ‘doers,’ they must keep dreaming to ensure their initiatives remain forward-thinking and innovative.

HFS, in partnership with Tech Mahindra, recently hosted a roundtable session in London. We brought together a group of GenAI ‘doers,’ including representatives from the BT Group, Citi, GSK, and Unilever. We discussed the reality of GenAI, how enterprises can move beyond dreaming to start doing with this powerful technology, and how combining dreaming and doing is the key to GenAI’s short-, medium-, and long-term success.

Exhibit 1: Tech Mahindra CEO Mohit Joshi delivers his opening remarks

Source: HFS Research, 2024

GenAI’s short-term hype and long-term potential are not the same. It’s not as simple as to dream or do

There are two ends of the spectrum: people who believe GenAI will be more significant than the invention of the wheel and those who are much more skeptical. In reality, most enterprise leaders sit in the middle.

Take the example of the chief procurement officer for a $20B financial services firm who drank the Kool-Aid and ditched his traditional procurement suite overnight for a GenAI-enabled autonomous sourcing solution. Compare it to the chief operations officer for a major P&C insurer who spent two years piloting three solutions for unstructured broker submission data ingestion. She remained sufficiently skeptical and only signed a one-year deal to see how the chosen solution worked. There must be a place between jumping in headfirst with wild abandon and taking three years to make a single decision.

The divergence comes as many executives are swept up in the excitement of GenAI’s hype and how it can transform their organization today—much of which is unrealistic and exaggerated. However, GenAI’s long-term transformative potential is very real. Enterprise leaders at our roundtable agreed that GenAI is overhyped in the short term but underhyped in the long term.

We surveyed the executives attending the roundtable and found that almost three-quarters believed they are currently GenAI ‘doers’ (see Exhibit 2). Throughout our conversations, it was revealed that nearly every organization represented in the room met our parameters to be considered a GenAI doer. Their attendance alone was likely indicative of their focus on the technology. Their GenAI maturity means their perspectives are unique—these are executives who are past the navel-gazing of GenAI dreamers.

Exhibit 2: 73% of attendees ranked themselves as a GenAI ‘doer,’ compared to only 35% of the larger sample

Source: HFS Research, 2024

Executive attendees agreed enterprises should adopt an approach to GenAI that combines dreaming and doing. They must first overcome the initial hurdles to become a GenAI ‘doer’—which we outline in the next section—while maintaining a long-term vision for the future of GenAI in their organization. Simply put, enterprises should ‘do’ today but continue to dream about the future. Mohit Joshi explained that approach:

We need to be impatient with the actions but patient with the results.

– Mohit Joshi, CEO, Tech Mahindra

Enterprises’ GenAI journey should start with “no regret” use cases

Too often, we see enterprises overlooking smaller use cases because they expect minimal returns—they are typically dreamers. However, small efficiency gains shouldn’t be forgotten; they compound over time, allow enterprises to experiment with GenAI, and build a culture of change within their organization. Kimberly-Clark’s Jim Edwards brought this idea to the group:

Easy use cases are overlooked because the prize isn’t big enough, but small things can reap massive benefits. A 1 or 2% efficiency increase in the supply chain is huge. Go after the big fish, but a school of small fish can be impactful, too.

– Jim Edwards, Global Innovation Capabilities Leader, Kimberly-Clark

Ultimately, there are more GenAI opportunities than bandwidth to make the changes. That’s why enterprises should focus today on the “no regret” use cases—the small fish—while maintaining a long-term vision. For example, layering a GenAI-powered solution over a company’s FP&A process enables the FP&A team to identify key projects within projects that are embedded in PowerPoint and ensure they are incorporated in the fiscal year and 3-year plans. Controllers can’t lose when they have additional insight. That’s how you dream and do—and it’s what the executives in the room said they have already done.

Leaders came to the roundtable armed with examples of how their organizations are leveraging GenAI today to help their peers identify their own “no-regret” use cases. The group widely agreed that efficiency-driving use cases were often the best places to start rather than those that might drive a new competitive advantage in the market. Examples included demand forecasting, sentiment analysis, and even drug discovery.

Enterprises share the same concerns and challenges about GenAI adoption

Every organization has unique circumstances, but we identified a handful of widely applicable concerns and challenges when asking leaders about their GenAI roadblocks. These include:

• Sustainability: GenAI’s environmental impact is often underestimated despite the extremely high carbon footprint of training a single GenAI model. Enterprises must be aware of the sustainability aspect of their GenAI investments, or the regulators will catch them off-guard.
• IP ownership: A lack of clarity on who owns GenAI inputs and outputs has left many enterprises hesitant to scale their GenAI projects due to concerns about risk exposure. We’ve already seen OpenAI and Stability AI in the courtroom with the New York Times and Getty Images, respectively.
• Hallucination: Enterprises are still facing the challenges of hallucination, which can lead to errors, misinformation, and an overwhelming lack of trust in GenAI. Despite this, many enterprises have yet to define a clear strategy to identify and address hallucinations.
• Talent: Access to talent remains one of the biggest challenges to any digital engagement—in particular, there is a lack of people who understand both the technology and the business. Additionally, their existing workforce must be willing to embrace change. A handful of executives told us that despite investing in Copilot licenses, employees weren’t using them!

The Bottom Line: Become a GenAI ‘doer’ first, but the key to unlocking the technology’s long-term potential lies in dreaming, too.

Enterprises exploring GenAI must move beyond dreaming to doing—it’s that simple, and we have already published the blueprint. It starts by pushing forward with no-regret use cases. However, the key to long-term GenAI success is creating a strategy and vision for the technology, including overcoming the abovementioned hurdles. Enterprises will likely see GenAI success across the short, medium, and long term if they correctly balance doing and dreaming.