CTO’s Guide: Creating Business value through GenAI
Sreedhar Gade
Vice President, Engineering & Responsible AI at Freshworks | The Customer & Employee Delight Engine
2023 has been an year thats been challenging the CXOs across industries to explore, innovate, differentiate through Generative AI. To be honest, below questions have been on minds of senior leaders all through this year:
In this article, I made an attempt to share my thoughts at 10K feet overview on what could a CTO could focus on during initial part of the journey, and will share more practical aspects in subsequent articles in this series.
1. Defining the purpose & Scope
Since GenAI is at the centre of the disruption currently, its natural to start white-boarding with GenAI phenomena and think about what we can do with GenAI to be differentiated in the market.
Recommendation is to switch from Traditional thought process of "What can we do with GenAI in our company to differentiate ourselves", on to "to what use cases of my business would GenAI add value by improving, accelerating or transforming". If we think about it, this is how most complex problems were solved in the past and this no different.
Have your GTM, Product, Customer Success & Leadership teams to list ALL the use cases that are built today and will be built tomorrow (1-2 years) using traditional software engineering, and brainstorm on which of these subset of use cases be built by GenAI to accelerate customer value & product' differentiation. This is an iterative process and takes multiple passes. Also, please remember that this is experimental. Be willing to shelve/dump some of the features that don't attract customer adoption and found to be not that useful. Fail Forward.
2. Freedom the choice with LLMs - Model Gardens
Sticking one Large Language Model or LLM provider would limit effectiveness of use cases and there by the business value. Build or buy a Model Garden/Contain like we see with AWS Bedrock, Azure ML Libraries, Google Vertex or a custom built one which is being built by the technology AI pioneers.
Will share this more in practical detail in subsequent articles, but pls do know that you need to create freedom of choice for your developer community. Gone are the days where we work tirelessly with a coding language to bring out a desired functionality. Now its era of finding which model can do with what level of tuning, than bringing in a generic one and invest huge amount of engineering investment to customize - For all you may know and most likely, problem you are trying to solve might have been solved by another model - We just have to look for it, test and tune...
3. Crafting differentiated features
Here's where big brains of the engineering & product teams be put for the use. We tested & selected an apt model in step #2 above, now its time to build a feature that WOW's the customers. Something that solves their chronic challenge or tranforms the way the employees or customers of your customers do their business.
领英推荐
Crafting a GenAI feature needs internal & external research, iterative approach with models testing & private beta phase, until you get to a stage that delights the customers and differentiates your products across your industry....
This will call for additional R&D budgets in short to medium term with exponentioal ROI if invested right.
4. Marketing, Pricing & Revenue Generation
What do we tell the community at large: A GenAI posture, mindset & policy needs to be announced so that your customers, investors and critics know that you are in the game!
Build a dedicated marketing strategy around showcasing how your company is building products & engineering abilities in GenAI field and how you are planning to disrupt the industry using GenAI. A series of posts, events, talks & policies to be published at repeat cadence to keep up the excitement in the marketplace and support needed for your GTM teams.
GenAI Feature Pricing: Leverage beta deployments at no-cost to validate effectiveness of features & market sentiment time to time before charging the customers. This is a new fields, its matter of time a competitor finds equal or better way to solve a problem and you don't want get into situation where you will have slash prices.
Build the pricing strategy based on business value realized by the customers rather than GenAI investments & cost incurred. This will stress your Gross Margins in short term considering rising GenAI Capex/Opex, however, its the best way to build long term & sustaining market credibility & revenues.
GenAI Pricing strategy is a large subject, but in summary, you will have to use a combination of 1. Blanket GenAI package (Annual/Bi-Annual) covering a group of key features 2. Premium add-on features like added governance, security or functionality.
Pivot your pricing on uniqueness that your company has, for ex: Vast amount of data pertaining to CRM domain, running GenAI on which will create unique value proposition for the customers. This would offer a combination of AI feature + unique data set your company possess which only a limited set of companies in your domain can offer. This is an combination that your customers can't refuse!
GenAI Revenues:
GenAI is a field with immense potential, but very little clarity on how the potential could be derived at this point. Clarity is evolving and will evolve at exponential pace. And hence,
have longer term mindset by not planning for immediate revenue generation and instead invest more on additional R&D budget for fail-forward and iterative approach to test out new use cases & features at fast pace.
Stay tuned in for further articles for CTO's GenAI arsenal for practical, real-life scenarios on how each of above steps can be deployed in businesses.
Hope this was helpful! Wishing you success in building & rolling out GenAI features that will delight your customers.
AI Recruitment Leader for GCC Unicorns in India. Delivering EVP, Talent Analytics, Talent Branding, and DEI Excellence.
1 年Well articulated