“Anxiety or Enthusiasm?” – asked Emily Chang at the Bloomberg Technology Summit.
“You need to have both – the thoughtfulness, the understanding, the nuance, and the tension between the two exists everywhere,” said Sam Altman, sharing his experience after travelling the Far East, speaking to users, developers and world leaders.
Altman’s observation pretty much sums up Generative AI adoption in enterprises, globally. It has been creating quite a bit of stir for enterprises, fuelling both enthusiasm and anxiety.
AIM got in touch with technology heads and CXOs of leading Companies across industries, alongside tracking the trends, to understand their adoption strategies, and the answers were quite surprising. Unlike previous AI adoption, which were mostly top-down, generative AI adoption seems to be being pushed in all the directions – i.e. top-down as well as bottom-up.
Bottom-Up Approach
Ironically, this is also the first time in the history of AI adoption where the bottom-up approach is gaining immense traction, where employees and teams come together, identify opportunities with use cases and PoCs, and are later supported by top management.
This is mostly driven by enthusiasm.
Most of the startups and growing companies fall under this category. These organisations are mostly experimenting with generative AI to address needs that benefit the company and their customers. Hopefully, along the way it might improve efficiency and increase productivity. Some of the examples include Swiggy, MakeMyTrip, Tech Mahindra and others.
Amitkumar Banka, head growth marketing at Swiggy told AIM that the company is using generative AI to create customised food images based on specific requirements on their platform, and this is helping them serve millions of customers.
“From a Swiggy perspective, it is a bottoms up approach. Most teams, including analytics, product, design, corporate strategy have come together to form a strong point of view in terms of how Swiggy should take generative AI to the next level. Each person and team are coming up with their own use cases to take advantage of generative AI.”
Narasimha Medeme, VP head data science at MakeMyTrip, have already launched conversational bots for customers and are now working on other features such as speech-to-text, and including Indian languages in their systems.
Top-Down Approach
Top-down approach has always been a go-to strategy for most companies, as it is faster and easier to consult and adopt. For instance, of late, a lot of IT companies – the likes of TCS, Infosys, Wipro, Cognizant, Accenture, and others, are partnering with Microsoft and Google to unleash their generative AI initiatives, alongside other technology enablers like Oracle, SAP, Salesforce and Zoho.
The adoption trickles down from the top. Qlik SVP Geoff Thomas believes that if a company wants to adopt a data culture and become a data-driven company, it would require strong sponsorship and support from the highest level of leadership. “It is often driven from the CEO, from top to bottom.”
But, there is a flipside to this, this might be exciting for people on the top or leadership team to improve their efficiency and productivity, but it often fuels anxiety among employees and teams, particularly those who have been familiar with traditional methods and productivity tools.
This also requires additional push from the organisation to offer training and certification programmes. Recently, Infosys announced a comprehensive and free AI certification training programme.
Hybrid Approach
Here, most companies are following both top-down and bottom up approaches, alongside setting up a centre of excellence to fuel generative AI use cases. Some of the examples include HCL Tech, Infosys, Zoho, and others.
Infosys recently unveiled Topaz which is a set of solutions and platforms using generative technologies with 12000 AI use cases, and 150+ pretrained AI models.
Wipro is also dwelling on a hybrid route. The company has partnered with Google Cloud to leverage its generative AI tools. Wipro will also integrate them with their own AI models, and pre-built industry solutions.
So, What’s the Best Approach?
In conversation with AIM, Ramprakash Ramamoorthy, Director – AI Research, Zoho Corp. vouches for hybrid approach, where he believes that the right mix of narrow and large models will be a win-win for companies and their customers.
“Narrow AI, where one model is trained to do one task based on past experiences will thrive, helping companies automate redundant tasks. Whereas, LLM-based generative AI will augment these capabilities by offering seamless availability of information from across sources.”
Zoho has a suite of 13 applications that is integrated with ChatGPT. “ Our focus right now will be to tightly integrate our AI stack across our product suites and also, in parallel, build in-house LLMs for businesses to provide seamless user experience in our offerings,” said Ramprakash.
Each industry has its way of implementing generative AI in their functions. Deepika Kaushal, Deputy Vice President at Piramal Finance, confirmed to AIM that their company is still identifying use cases for generative AI applications and it is better for them to learn from the experts and later in future learn the capabilities to build in-house.
On the other hand, Vivek Sahabadi, Head of Data Analytics at Navi, is of the opinion that the kind of data a company handles decides the right approach. “For fintech companies, where user data is critical, building their own models makes sense, whereas, in industries such as food tech, external models can be used.”
All in all, it becomes important for companies to consider factors such as costing, expertise, data security, and in-house infra capabilities, before diving into the adoption of generative AI. Most importantly, the usefulness or second order understanding of generative AI should be established. One must not recklessly rush into it, irrespective of whether excitement or anxiety is pushing them towards it.
The post How Companies are Botching Generative AI appeared first on Analytics India Magazine.
