What’s the Hold Up On GenAI?

(Overearth/Shutterstock)

When generative AI landed on the scene two years ago, it was clear the impact would be sizable. However, the path to GenAI adoption has not been without its challenges. From budgeting and tools to finding an ROI, organizations are figuring out as they go along how to fit GenAI in.

Here are 10 questions about the GenAI rollout and how it will impact your business.

1. What’s the GenAI budget?

In the overall IT budget, AI will be a significant portion of any new or fresh funds that the business allocates for spending. In terms of use cases, the largest share of the Gen AI budget is likely to support applications such as implementing chatbots, getting data from knowledge bases into other conversational content platforms. The goal for this budget will be how to enhance user interaction, streamline information access, and improve support and engagement through conversational AI interfaces.

2. What is the current state of generative AI in production across industries?

Generative AI is still in its early stages of adoption, with most businesses yet to launch their first production-grade applications. While tools like ChatGPT demonstrate potential, the reality is that widespread deployment—especially for business-specific use cases within enterprises—hasn’t occurred. The delay mirrors previous technological waves, where enterprises took between two and four years to integrate new innovations meaningfully.

So, 2025 should be the year when we see companies actually launch and have to make good on their promises around AI, both internally and to the market. Those companies that do this successfully will see huge market impact.

Chatbots are step one in the GenAI adoption curve (sdecoret/Shutterstock)

3. Why do some experts criticize the “more than a chatbot” narrative?

The “more than a chatbot” narrative is seen as premature because most organizations haven’t successfully implemented even basic chatbot systems that deliver on their promises to users. Many IT leaders and vendors who advocate for more advanced applications often lack experience with actual chatbot deployments. Getting the right foundations in place is essential, and that work on GenAI projects should not be devalued in the rush to hype the next big thing in AI.

4. How does the adoption of generative AI compare to previous technological shifts like mobile and social?

Generative AI adoption is following a similar trajectory to previous innovations like mobile apps and social media. Look at mobile – Apple launched the App Store in 2008, and it took to 2009 for Uber to launch and 2010 for Instagram to launch their apps. Each of these apps disrupted industries . For example, Mobile enabled Spotify to disrupt the music industry and Airbnb and Uber disrupted the hospitality and transportation industries. Those companies are now worth billions. It took even longer for traditional enterprises to feel comfortable with mobile, yet now it is essential to them. GenAI is following that same path, and we are now in that two year timeframe. So we should see some strong launches in 2025 and beyond.

When ChatGPT launched, it was impressive to a lot of people. But Gen AI needed development tools around it, and around the other LLM tools that launched after, in order to become something that enterprises could take and use at scale. It needed approaches like vector data embeddings, vector search, integrations, and all those other elements that go into making technology work at scale. Those tools are getting into place, and 2025 should be the year when those deployments start coming through.

5. What are the challenges facing businesses in deploying generative AI?

There are four key problems – inertia in adoption, lack of expertise, getting over the hype and having the right infrastructure in place and ready. Many enterprises are slow to experiment and deploy new technologies, even when they are production-ready. GenAI is still developing, so there’s a lot of companies that are still adopting a wait and see mindset. But GenAI works best when you use your own data with it, so you can’t copy another company’s approach and expect to get the same results.

The difficulty of finding GenAI developers is hindering adoption (Gorodenkoff/Shutterstock)

Linked to this there is a lack of expertise around GenAI out there–finding the right people that can manage and scale AI deployments is hard, simply because the number of people out there is small.

The amount of hype around GenAI is not helping this process either. A lot of what we use as inspiration for how we think AI will develop is found in science fiction, and that fiction has led to some unrealistic expectations. The gap between what Gen AI can deliver today and how it can be used in practical business applications leads to delayed implementations. We have to temper expectations and concentrate on real world environments where we can compare ‘before and after’ results.

To be ready for GenAI, businesses need better tooling, architecture, and observability systems to integrate AI solutions effectively. The large language models have attracted the majority of attention, but they are only part of the approach. You can’t deliver Gen AI without the right data, the right tooling, and the right information around how you are performing.

6. What industries are expected to benefit most from generative AI?

Industries that rely heavily on engagement—like customer service, retail, and support functions—are poised to see the most immediate benefits. As well as industries that are limited by cognitive burnout of highly specialized people. AI-powered tools can enhance customer interactions, improve support efficiency, and provide real-time advice for field operations. More specifically, AI-powered tools can enhance reviewing medical scans, delivering highly technical features and drug discovery. However, achieving these benefits depends on overcoming deployment bottlenecks.

7. What is the role of venture capital in generative AI, and what mistakes have been made?

Venture capital has played a significant role in funding generative AI, but many firms overemphasized investments in model development rather than broader AI infrastructure. The value in generative AI lies more in software applications, tooling, and orchestration than in training new models. VCs are shifting focus toward infrastructure and deployment solutions, but many of these firms lack experience and expertise in the B2B software sector. They don’t understand the buying patterns that large enterprises have, and this will affect how those companies that got funding will perform over the next year.

GenAI startups are attracting billions in venture funding (TSViPhoto/Shutterstock)

I expect there will be companies that have great parts of the stack, but they don’t have the funding to get to market effectively and scale up. This will lead to a lot of mergers, acquisitions and financial opportunities for those companies that are able to get a strong position in the market.

8. What predictions exist for the future of generative AI adoption?

2025 will be the year where we go from hype to widespread production use and deployments around AI-powered chat services or where AI gets embedded into other applications. We’ll get where we’re going faster. For Scientists, generative AI is going to reduce the cognitive burden of scientists globally and the world will be a better place for it. For technologists, generative AI will build products faster, fix bugs when we find them, and deliver experiences users love. We’ll get where we’re going faster, we’ll cure cancer faster, and we’ll combat hunger faster, with the power of generative AI in 2025.

Alongside this, I think the research side will continue to develop rapidly. Over the next year, we’ll see new terminologies and concepts emerge, even as many businesses are still catching up on deploying current technologies like chatbots. This will help more complex deployments to get completed, and then expand what Gen AI can deliver.

9. Why are current chatbot use cases still relevant for 2024 and beyond?

Although conversational interfaces (chatbots) might seem like “last year’s use case,” most organizations haven’t implemented and deployed even one in production effectively. Therefore, deploying conversational interfaces remains a critical goal for 2024. For enterprises, the emphasis is on creating functional and scalable solutions for customer interactions, internal support, and field operations.

10. What is the long-term outlook for generative AI in enterprise use?

Generative AI will likely become the fourth major wave of digital engagement after web, social, and mobile. Over the next few years, it will transition from an experimental technology to a core component of business operations. Companies that embrace generative AI to enhance engagement and efficiency will gain a competitive edge. For any area where enterprises can see more opportunity than risk, there are gains to be realized from GenAI. Unobtrusive LLM-augmented Assistants, not just in chatbots, but in understanding our world based on our digital exhaust. They become a copilot for life, advising on balls humans drops, handling the complexity of balancing work and life, stopping you from sending that flaming reactive email.

An agentic world can empower stakeholders to measure the right things about their business, change those measurements more quickly, and provide the critical perspective on whether the right decisions are being made for the business or enterprise. Imagine an executive working with their GenAI Assistant: One of our KPI’s is dipping. Help me figure that out. The chatbot says “Okay. based on what this KPI represents and the data available for analysis, I have three hypotheses”. AI agents could then test the hypotheses.

About the author: Ed Anuff is the chief product officer at DataStax, provider of a big data platform. Ed has more than 30 years experience as a product and technology leader at companies such as Google, Apigee, Six Apart, Vignette, Epicentric, and Wired. He led products and strategy for Apigee through the Apigee IPO and acquisition by Google. He was the founder of enterprise portal leader Epicentric, which was acquired by Vignette. In the 90s, at Wired, he launched one of the first Internet search engines, HotBot, and he authored one of the first textbooks on the Java programming language. Ed is a graduate of Rensselaer Polytechnic Institute (RPI).

Related Items:

Focus on the Fundamentals for GenAI Success

GenAI Begins Journey Into Trough of Disillusionment

GenAI Adoption: Show Me the Numbers