Scanning the Big 4 Gen AI Horizon (A Primer)

Image generated by DALL*E

Talking to people about Gen AI is like walking through a flea market with cash in your hand. Everyone wants to sell you on their product or promise. You’re in the market for a reason. Where do you start? 

How do you know who in the marketplace is going to help or hinder your efforts?

There are amazing accomplishments in AI and Gen AI around nearly every corner. If you haven’t seen ChatGPT’s Sora and the ability to create life-like video from text, well, just keep your head in the sand. There are lists and lists of the Top Generative AI companies cropping up almost weekly. Of course, there’s always the GenAI landscape charts like this one from Datacamp (which are quite helpful in many cases where you need information at a glance). Not all of these are relevant to your goals, strategies, and desires. 

You don’t want to choose between all Gen AI purveyors, you need someone to help filter the marketplace down to the strategic, applicable few that will maximize your chances for success in each proof-of-concept. You want to win.

Boston Consulting Group laid out five characteristics that set apart “winners from observers” in their January 2024 article, “From Potential to Profit with GenAI:” 

  • Investment in productivity and topline growth; 
  • Systematic upskilling; 
  • Vigilance about AI cost of use; 
  • Focus on building strategic relationships; and 
  • Implementation of responsible AI principles.

The key thing to note which goes unsaid is that accomplishing these five things requires a focus on people and process, not the tech itself. Productivity is a measure of human efficiency. It requires people to adhere to and improve in using a process. The ad hoc application of GenAI will decrease your productivity if you do not get colleagues to abide by the process improvements. 

The other four characteristics are even more focused on people and leadership. Upskilling requires having training and support programs to develop skills. Cost capture, through methods like activity based costing, is more about people’s ability to track and manage tasks accurately to ensure the right allocation of resources. You can’t claim cost savings you don’t capture while simultaneously increasing the cost of compute. Building relationships and responsible principles both hinge on your people. You need leaders and people in other parts of your business to be part of this journey.

Every technology shift requires adoption by people. Successful leaders create moments that help people make the change. 

McKinsey & Company provide a reminder in this article that, “if your data isn’t ready for generative AI, your business isn’t ready for generative AI,” and the role of the Chief Data Officer is, “to be clear about where the value is and what data is needed to deliver it.” The portfolio view and framework here are a good reference for anyone looking to provide AI use cases to the C-suite. 

Bain and Company take a broader approach. Rather than providing insights bespoke to GenAI, they look at the application within industries and departments. The Bain Insights blog on Artificial Intelligence is a good resource for thinking about the impact on topics ranging from healthcare and retail to procurement and finance. 

Accenture and IBM take the informative route. Each of these large firms provide a good overview to topics in Gen AI balancing technical depth against industry nuance. Accenture’s Generative AI Insights is definitely geared to marketing, but it provides meaningful definition and context if you’re unsure about certain jargon. IBM’s Artificial Intelligence blog may be the most informative. It provides you with the spectrum of AI from the classic, machine learning beginnings all the way to “Theory of Mind AI” and AGI of current science fiction. 

There’s a lot of information piling up about AI. Nearly every company has a product infused with AI, and everyone has a perspective about the benefits. Only you know what you’re looking for in the marketplace. If you make a moment to hone your desires you can position the right next step toward success.

When you’re ready to create that moment, we should talk.

Set the Tone Regarding AI

Image by Freepik

Let’s say the quiet part out loud.

Implementing a revenue-generating, production-level AI solution is a bridge too far for many companies right now. More specifically, implementing Generative Artificial Intelligence (Gen AI) is something that extremely few enterprises are fully prepared to take on. Recent research published in Harvard Business Review and Forbes show nearly 80% of AI projects fail if you don’t have your data fundamentals under control. On the MIT Sloan podcast, “Me, Myself, and AI,” co-hosts Sam Ransbotham (Boston College) and Shervin Khodabandeh (BCG) focus on the fact that only 10% of companies succeed with AI.

With something this significant, why are failure rates so high?

For one thing, take a look at the portfolio of most C-level technology leaders. Across sectors, tech leaders are tasked with digital transformation which uproots the business model, recruiting within tight labor markets that create attract-retain staffing issues, engaging “return to work” vs “remote-first” policies that challenge networks and collaboration tools, supporting various modes of enterprise systems across cloud and on-premise infrastructure, launching data science products to gain customer insight, and keeping all of these attack surfaces secure from cyber threats. Shall we continue?

Digital and Technology leaders struggle to parse signals from noise.

Everywhere you turn someone’s going to be extolling the need for you to get your AI strategy and use cases lined up. A recent survey from Boston Consulting Group showed that 85% of executives are going to increase spending on AI and 89% consider Gen AI a “top 3 priority” for FY24. Software and service vendors alike want you to spend on AI. It’s not that you can’t start exploring. Effective leaders need to cut through the noise creating distractions and communicate a consistent vision that aligns people and outcomes.

Companies that are analytic competitors see the most gains as first movers on Gen AI use cases. They’ve made the investment in data management, statistics, and math for years and know what data serves as foundation for Gen AI success. There aren’t many ways to circumvent data when launching AI projects. If you don’t have this foundation set, your next step is clear.

Think of it another way: the AI/ML platform Hugging Face hosts over 350,000 ML models. Each model can contain hundreds of data inputs or outputs. We are moving away from data management and toward model management. If you do not have strong practices in place for strategic data assets, your AI initiatives will suffer or likely fail. Many companies are still placing people and practices around strategic information management.

So where do you start your AI roadmap? Set the tone.

Questions will abound. What is acceptable use? Are you looking to understand the implications of Gen AI in your market or are you wanting to enrich a competitive advantage of your firm? As a leader, what is the one thing you need to gain during this stage of the AI hype cycle? What is the near-term roadmap for Gen AI initiatives being considered? Who is accountable? 

Cut through the hype and hyperbole to make meaningful progress. Articulate a compelling statement that clearly demonstrates why you’re making the investment, where you’re focusing on the business strategy, what you’re expecting as acceptable behaviors, and how you’re measuring the return. Setting the priorities, policies, and practices regarding the strategic direction of AI, provides signals for managing expectations of people both in and outside the company. 

The era of AI is a battle for trust. Competing with Gen AI is not about chatbots and avatars. It’s about setting the edge where customers engage with your company in a digital ecosystem. Brand loyalty is tied to how customers perceive what your firm is doing with Gen AI. Employee attraction and retention ties-in to ethics in your hiring practices. If people distrust you they will disown you. 

Get loud about what AI is and is not for your business goals. Clarify the strategic purpose of AI by creating a clear voice for the company, C-suite, and Board. Let your outcomes evolve with the market. Once you’re positioned to be heard, you’ll be ready to take on any AI challenge.

Want to talk about crafting your voice?

We have a problem

Image by master1305 on Freepik

More specifically, it’s likely that you have a problem.

Mulesoft, a Salesforce subsidiary, recently published their 2024 Connectivity Benchmark Report which surveys over 1,000 IT decision-makers and the findings are alarming. Consider that 98% of IT leaders report digital transformation challenges holding their companies back. Of those leaders, 81% state that data silos are key shortcomings of the transformation. 

We are one year into the era of Generative AI – an era predicated on the trust and truth of data – and yet almost every technology leader is unable to unlock digital value for their companies because of their inability to manage and harness data. Sound familiar?

Big data was a thing more than a decade ago. Amazon, Microsoft, and Google brought us to the cloud in 2006. Research firm Gartner forecasts that IT budgets will grow nearly 7% in FY2024  as business departments must rely on automation to innovate. Here’s our problem: most IT leaders have not solved the fundamental problem of using data to drive transformation. If your systems and applications are still being built in ways that perpetuate silos, increase rework, or require bespoke teams to enable use by the business, something’s wrong.

It is staggering to think that 98% of technology leaders struggle with digital transformation. It is equally concerning that data are a constraint in 2024. There are two questions to ask: why haven’t companies liberated their data (created processes and standards that facilitate data), and since these silos persist, how are those companies going to leverage AI (without creating yet another silo)?

Here’s the good news.

Breaking the data silos enables all of these ambitions. It’s the most impactful “two-fer” that you can give your company. Companies leading the marketplace in digital transformation, revenue growth, and now the GenAI horizon are analytic competitors – “firms that compete on the basis of their mathematical, statistical, and data management prowess.” IT leaders seem to have glossed over the nuanced definition given by Tom Davenport and Jeanne Harris seven years ago. It is those three components taken together – enabling data management, doing the math, and acting on the statistics that create value and win.

So when it comes to enabling transformation are you in the 2%? If not, we should talk.

Accelerating Integration Insights

In the last three decades, there has been an tremendous growth in the area of information processing needs of data-driven businesses in government, science, and private industry. Businesses compete on capturing, integrating, and analyzing data to help knowledge workers make sound business decisions and create growth. Data integration across companies (legal entities) is one of the most challenging steps.

To date, three (3) key levels of data integration have been identified across enterprises. First is data access. During most merger & acquisition (M&A) activities, Corporate Development Officers settle for having access to data from each enterprise. All of the work goes into finding facets of data that relate to each other, supporting the same grain of analysis, and finding people knowledgeable enough to analyze each set. The manual and duplicative effort creates its own complications and delays to providing accurate insights.

Second, is pulling data into a common platform. Some companies are able to pull data into one of the platforms for subsequent analyses. While not integrated, this level of automation and access allows a single team to analyze data relevant to all companies in the portfolio with a common set of basic business rules.

The final model is nirvana, the consolidated data model. At this point, the data from each company augments each other. Executives can examine the impact of one company on another, time series analyses, forecasts, and pricing analytics. This level is rarely achieved early on in the lifecycle.

Each level of integration has its own set of complexities that requires a certain amount of time, budget, and resources to implement.  

We created a methodology based on industry best practices to measure the readiness of an organization and its datasets against the different levels of data integration. Our Integration Level Model (ILM) and ILM tool are used to quantify an organization’s readiness to share data at a certain level of integration. It is based on the established and accepted framework provided in the Data Management Association (DAMA-DMBOK). It comprises several key data management functions and supporting activities, together with several environmental elements that describe and apply to each function. By scoring the maturity of a company’s systems and data in a pragmatic method, we can target specific areas of an integration to prioritize. This reduces the complexity, cost, and risk associated with multi-enterprise data integration, and accelerates the time-to-value of analytic insights.

https://www.researchgate.net/publication/269321702_A_qualitative_readiness-requirements_assessment_model_for_enterprise_big-data_infrastructure_investment

https://www.ornl.gov/publication/qualitative-readiness-requirements-assessment-model-enterprise-big-data-infrastructure