Suggestions
Aakash Dharmadhikari
Building the AI Transformation Platform for Salesforce Partners | ex-CPO - Jiva, Gojek Group
Hi, I'm Aakash, the co-founder of realfast. At realfast, we are building the AI Transformation Platform for Salesforce Partners.
My last startup, C42 Engineering, was acquired in 2015 by Gojek Indonesia, one of the fastest-growing startups in the world and Indonesia's first unicorn, growing from ~500 orders/day to ~7 million orders/day and $10 Billion in GMV in just 4 years. Gojek then went on to merge with Indonesian e-commerce powerhouse Tokopedia to form the GoTo Group, which went public in 2022 at a valuation of ~$30 Billion.
At Gojek, I was the Chief Product Officer for GOJEK group, served on on the Board of Directors, and played several key roles over the years, including SVP Product Analytics.
After that, Aakash served as the Chief Product Officer of Jiva (https://www.jiva.ag/). Jiva is the largest AgriTech Startup, serving Indonesian Small Holder Farmers - from credit to buy seeds to buying the harvest at the farmgate.
Highlights
Unpopular opinion: The idea of a single, monolithic AI agent that can handle any task or workflow thrown at it is more science fiction than practical reality.
I know it's appealing to think that way. But it's far from the truth.
At realfast, what we've discovered over the past year is that effective AI solutions need to be tailored to specific use cases and individual workflows within them, deeply understanding the intricacies and requirements of each workflow.
Especially in IT services automation, don't think there can be a one-size-fits-all AI agent. Each workflow has its own unique challenges, dependencies, and success criteria. An AI agent that automates network provisioning will need a very different set of capabilities and domain knowledge than one that handles incident management. Also, the data sources, business logic, and performance metrics will vary significantly from one use case to another.
For example, at realfast, we are currently building an AI-powered unit testing agent for Salesforce Partners.
Each implementation will be an AI agent specifically trained for unit testing on Apex code, and customized to every client's business use case — their custom objects, fields, validation rules, automations, and integrations.
The agent is trained on Salesforce best practices for Apex testing. It will analyze the structure and dependencies of Apex classes, generate relevant test data, and create optimized test cases that maximize code coverage and edge case handling.
Now, the goal is to build a highly specific unit testing agent that acts like a junior developer working together with the QA team to ensure comprehensive coverage, catch critical defects, and enable faster, more reliable delivery.
And we have had to collaborate with multiple Salesforce partners and use their expertise to build this.
So, the idea of a general-purpose AI agent that can magically handle all possible development scenarios is a pipe dream.
Otoh, leveraging AI to automate and optimize specific workflows is not only possible today, it is increasingly going to automate a huge chunk of IT services.
