AI has been talked about everywhere and is being incorporated everywhere, but it is often embedded in places many people do not even realize or cannot see. Generative AI was such a sea change because it gave the public AI that they could easily touch, feel, and interact with.
When ChatGPT launched in late 2022, it felt like magic. It unleashed an even more tremendous appetite and fascination with AI technology.
Marketing and advertising caught wind of that fascination and billed AI as a panacea that could deliver magical results. Big promises and outrageous predictions were made.
Despite the promise, AI is not magic—however, I believe it is inevitable. We are in the confusing phase of AI evolution where we cannot yet see exactly how it is going to progress. Like any disruptive technology, there is infinite potential.
LOOKING TO PAST TRANSFORMATION
We know from previous transformations that it takes time to fully integrate new technology. Cell phones and laptops evolved tremendously over time and became seamlessly integrated into the foundation of modern life and work.
AI is still in the early stage. It will be a long journey, but the transformation of the knowledge worker is inevitable. We are just starting to scratch the surface of what is possible.
TODAY’S CHALLENGES WITH ENTERPRISE AI
The challenge with AI right now is two-fold.
The first is that there is no infrastructure for enterprise AI. There is a layer of software that needs to be built between the model and the enterprise, and the tools are not there yet. Once that technological infrastructure is built, then AI will become real.
The second—and bigger problem for enterprises to solve to succeed in AI—is their data. Data, teams, and processes must be brought together to see the full benefits of AI.
The are many other real concerns that exist when it comes to AI like compliance and security fears, the notion of AI going rogue, etc. These will eventually fade, in my opinion. I believe the first adopters need to build the critical foundation and demonstrate successful use cases, then the pile-on will really accelerate.
ASSESSING YOUR DATA STRATEGY
For leaders looking to assess their current data strategies, processes, and AI readiness, there are some specific steps they can take:
• Consider where your CX data lives. Challenge yourself to think beyond clean, in-use customer experience data. If you had unlimited hours to read through and understand CX data, what holds the best insights? If possible, would you benefit from pulling away the noise that hides these insights from you?
• Consider how your CX data gets organized, structured, or labeled. Count how many teams employ manual, automated, or AI-powered data organization. Would you benefit from unifying this organization so all data is structured similarly?
• Consider who across your business benefits from having these CX insights at their fingertips. Stakeholders could include customer service agents, marketing teams, local customer-facing representatives, or those responsible for collecting customer feedback. Do these stakeholders span across more than a single team or business unit?
If any answer to the above questions is yes, your organization is likely prime for AI adoption.
PREPARING FOR AI-ENHANCED WORK ENVIRONMENTS
To prepare your team for the near future of AI-enhanced work, start by empowering them to learn and try AI within their day. This could mean sharing programs of top AI news outlets, premium subscriptions to tech publications, or simply solid guidelines on how to best try AI at work without putting the business at risk. This invitation can create positive momentum across your organization to embrace AI efforts.
I’d then challenge you to consider how you can leverage AI to activate your CX data (by pulling out unnecessary noise) in a unified manner, and organize it in a way that gives structure, so your teams across all customer-facing teams can act by deriving insight and driving better customer interactions.
MAGICAL RESULTS NOT INCLUDED
Most businesses are not yet seeing the results they expected from AI adoption in terms of AI’s ability to drive sales, reduce costs, improve CX, and mitigate risks. Many enterprises viewed AI as magic (or were sold it that way) and expected magical results.
The business results will come when organizations are able to implement and improve key AI—especially genAI. For traditional AI, the most important metric is accuracy, measured through an F1 score. Sometimes called F-score or F-measure, the F1 score combines precision and recall into a single score to evaluate the performance of an AI model.
For measuring gen AI, accuracy is important, but teams should also focus on monitoring:
1. Hallucinations: Like accuracy—how often is the content generated true?
2. Context alignment: Is the output relevant to the task at hand?
3. Clarity and conciseness: Is the output easy to understand and take action on?
4. Seamless handshakes: How well does it integrate into other systems and human employees when more is needed?
5. Guardrails adherence: How well does it stay compliant with any required guardrails based on regulatory or legal requirements?
FINDING SUCCESS WITH AI
AI is inevitable, but I do not anticipate there being a single tipping point when we realize it is fully embedded and delivering successful results for companies, teams, and individuals. Instead, I believe there will be a series of things that put companies into two buckets of winners and losers.
Success will most likely come to those who have integrated AI into the fabric of their operations, aligning every initiative with core business objectives. And those who fail will probably still rely on a collection of standalone, disconnected AI tools that cannot achieve meaningful impact.
Data infrastructure is the cornerstone of this success. It could very well be the rocket fuel that unlocks the transformative potential of AI.