AI: Hype and Reality

Why AI investment alone is not creating business outcomes, and what enterprises need to change

There is so much hype about AI today. Enterprises are experimenting with AI, and many have already invested in a big way to create outcomes. Yet outcomes and adoption remain very minimal. 

Major investments are being made in data centers and AI tools. Without a clear path to business outcomes, AI risks becoming another major technology spend without real business innovation results. Even the founders of frontier AI model companies acknowledge that models can mislead. The question for enterprises is not whether AI is powerful. It is. The question is how to use it to create outcomes the business can trust and own. 

The reality: 25 years of complexity 

Over the past 25 years, enterprises invested heavily in many different tools and technologies. The so-called transformation journey left most of them with complex business and technology environments and siloed data. Business was not able to effectively innovate, engage and take ownership of the outcomes for growth, operational efficiency and tech debt reduction. 

Adding AI on top of this environment does not solve the problem. It adds another layer to it. 

Everybody in business, innovating 

AI is very powerful when it works with relevant data. The real shift comes when business itself innovates with it. 

When everybody in business, in the trenches, is focused on innovation in a business-facing environment, with a business innovation methodology, oriented to business assets and engineered through the SDLC, they create business assets that are high-quality, reusable with no duplication, cyber secure, compliant, auditable and traceable. This environment sits outside the current complex and siloed production environment. 

Business engages and takes ownership of each outcome as a business asset it can trust, AI-Ready with AI Governance. A fully automated, end-to-end Sandbox enforces an agile, engineering approach for every use case, which brings consistency, scale, and speed. Because each asset is reusable, every outcome becomes a head start for the next one rather than another silo. 

All of this is driven by business judgement. That is what allows enterprises to adopt AI effectively and realize outcomes substantially for growth, operational efficiency, and tech debt reduction. 

What it takes 

• Data as a Product: data with relevance to business innovation, business ownership, MDM enriched with entities and relationships every step of the way, and AI-Ready AI Governance. 

• A business innovation methodology: an agile and engineering approach for each use case outcome. 

• A Sandbox: fully automated within the customer environment, business-facing, and outside the current complex production environment. 

• A path into production: business assets built through the SDLC that realize the impact in the current production environment. 

How it works: three steps 

Step 1: Leverage AI with a semantic layer in the Sandbox 

Bring data from various sources together centrally at the domain level. Cleanse it and create data products at the domain level, with relevance, business ownership, MDM every step of the way, and AI-Ready AI Governance. These become Data, AI, and ML intelligence assets in a central database or data mesh. 

The semantic layer is backed by an enterprise ontology: one governed model of the business that connects and contextualizes both structured data (ERP, CRM, transactional systems) and unstructured data (documents, emails, tickets, contracts, notes). It can also augment and modernize current applications at the domain level, making the business more knowledge-based. 

Step 2: Orchestrate bite-size use cases 

With these data products in place, business can orchestrate use cases at bite-size level. Even complex use cases are broken down into bite-size ones. Business innovates centrally to create Data, AI-Ready and AI-Governance assets that realize outcomes. 

Step 3: Fully automate end-to-end workflows 

Today, operational knowledge is carried by a patchwork of spreadsheets, email threads, standalone workflow tools and disconnected BI reports. With the AI-driven semantic layer, every stage of operation (intake, planning, execution, monitoring, reporting, and improvement) reads from and writes to the same governed model of the business, in fully automated, end-to-end workflows. 

The result is fewer hand-offs, one definition of every metric, and decisions made in the system of work rather than around it. This creates a very collaborative environment, brings very high productivity, reduces operational cost and eliminates technical debt substantially. It also makes the next innovation easier: new analytics, automation and AI use cases plug into concepts the business already understands, which drives adoption because users work in business language and get timely, relevant outcomes. 

Our journey 

We started with these concepts, including the business innovation methodology, 10 years ago. After four years developing them in detail, we started building the Sandbox in the customer cloud environment, for business, away from current silos and the complex production environment. It took another five years to fully mature it, experimenting with several use cases and engaging business to take ownership. For the last 1.5 years, we have been proving the impact through enterprise adoption and outcome creation. 

Methodology and technology also need the right talent. We created the AI Academy to develop business innovation talent through a six-month curriculum. FDEs, and practitioners with strong industry and domain expertise, solution the use cases, simplify engagement with business and engineering leaders, and orchestrate use cases in the Sandbox in an agile way. Together, they create business outcome assets for adoption and impact, leveraging AI to the fullest extent. 

We are also building an agent inference layer in the customer environment, using smaller LLMs, so business can query in natural language while complex enquiries are controlled to manage cost. 

Beyond the hype 

These are very exciting times for enterprises. The reality is that AI creates outcomes when everybody in business can innovate, when data is relevant and governed, and when there is a consistent methodology and environment in which to build trusted, reusable business assets. Enterprises that take this holistic approach will bring innovation for growth, operational efficiency, and substantial tech debt reduction.

Raj Vattikuti
Chairman and Founder, Calibo