Ravi’s Journey
Ravi started his career as a mechanical engineer in the nuclear industry. Ravi transitioned into technical sales in September 2014 selling mechanical hardware. He then broke into Software/Tech Sales in April 2017 when he joined MathWorks as an Inside Sales Engineer. Since then he has worked at several startups like DataRobot and WhyLabs, helping build out Solutions Engineering teams and gaining a wealth of experience in AI and machine learning applications.
Ways Ravi Dogfoods Akkio’s AI Functionality
- Lead Scoring: Ravi uses an Akkio feature called “Chat Data Prep” which employs natural language prompts to clean and transform data. They’ve plugged CRM data into Akkio to simplify complex data manipulation tasks that would otherwise require extensive coding in SQL or Python. Ravi mentioned, “We use Chat Data Prep to help with lead scoring. We’ve built a model based on historical data that’s always auto-updating, helping us prioritize leads more effectively.”
- Visualizations: Another powerful tool in Akkio’s arsenal is “Chat Explore.” This feature allows users to perform data analysis and generate visualizations through natural language prompts. It’s particularly useful for business intelligence and data analytics, offering an intuitive alternative to traditional coding methods. Ravi uses Chat Explore to help his team and his customers visualize ROI and quickly sift through their business data.
Ravi explains, “As opposed to writing pandas code or doing something in Excel or SQL, it's all natural language-driven. We leverage Azure OpenAI as our LLM, which generates the appropriate pandas code based on natural language prompts.”
1. Documentation
“I use ChatGPT to help me craft documentation in user-friendly terms, making it easier for our users to understand how to use our platform.”
Here’s how he does it:
- Drafting Product Documentation: Ravi is sometimes responsible for creating documentation for new Akkio features. ChatGPT assists in generating detailed product overviews, step-by-step instructions, and troubleshooting guides. This not only saves time but also ensures the documentation is polished and professional.
- Simplifying Documentation: Ravi often needs to explain complex technical processes in a way that’s easy to understand for non-technical stakeholders. Using ChatGPT, he can convert intricate technical jargon from documentation into clear, concise instructions. For example, if he needs to explain a new feature of Akkio’s platform, he inputs documentation into ChatGPT and asks it to generate a user-friendly guide.
- Writing Better Bug Reports: Another one of Ravi’s many responsibilities is filing bugs on behalf of his users. Ravi often takes descriptions of bugs from users, asks ChatGPT to provide a first-level diagnosis, and then uses that information to give more context to the engineering team. This helps his engineers troubleshoot faster, and improves response times for his customers.
2. Coding Assistance
“Even though I don’t code regularly, I can use ChatGPT to generate the necessary code snippets, debug errors, and optimize my scripts, saving me a lot of time,”
Ravi, although proficient in Python, does not code regularly. This is where AI steps in to bridge the gap:
- Generating Python Code: When Ravi needs to create or modify code, he uses ChatGPT to generate the necessary Python scripts. For instance, converting cURL and POST requests from documentation into Python code.
- Creating Custom Solutions: In scenarios where clients have unique needs, Ravi builds custom solutions using AI-generated code. For example, he created a custom Streamlit app around Akkio’s API deployment with the help of ChatGPT. This app included logic to check for certain conditions that weren’t covered by Akkio’s web app, enhancing the client’s experience and demonstrating the flexibility of their platform.
- Debugging and Optimization: AI assists Ravi in debugging and optimizing his code. If he encounters an error, he can input the error message into ChatGPT to get suggestions for resolving it. This is particularly useful for complex issues that might require a deep dive into documentation or forums like Stack Overflow.
3. Synthetic Data Generation
“I use ChatGPT to generate synthetic data for demos, especially when the data we need is hard to find. This makes our demos more relatable and impactful,”
Generating realistic data for demos and testing is another area where Ravi leverages AI:
- Tailoring Data to Specific Use Cases: When preparing for client demos, having data that closely resembles the client’s real-world scenarios can make a significant impact. Ravi uses ChatGPT to create synthetic datasets tailored to specific personas or industries. For example, if he’s demoing to a retail company, he can generate sales data that includes typical retail metrics and trends.
- Enhancing Predictive Models: While synthetic data is often not perfect for training predictive models, it’s excellent for showcasing exploratory data analysis and visualization capabilities. Ravi uses AI to generate datasets that highlight Akkio’s strengths in data preparation and transformation, making the demos more relatable and impactful.
- Creating Engaging Visuals: In addition to raw data, Ravi uses synthetic data to create engaging visualizations. This helps in demonstrating the full potential of Akkio’s platform during presentations, as clients can see tangible examples of how their data could be transformed and analyzed.
Encouraging AI Adoption Among Teams
While Ravi is a power user of AI, he acknowledges that not all his colleagues are as comfortable with these tools. He suggests that having a “point person” who is well-versed in AI can help others get up to speed. This individual can assist with prompt engineering and provide guidance on how to integrate AI into various workflows effectively.
Ravi shared a specific example of how he helps his team with AI:
“Our marketing team recently asked me to analyze call transcripts to understand better how prospects heard about us, what they liked about our product, and where we lost them during demos. Before LLMs, this would have taken days or weeks of manual work. Now, I can use ChatGPT to quickly parse these transcripts and extract valuable insights. I prepare the transcripts and use prompt engineering to have ChatGPT identify key points, which significantly speeds up our analysis process.”
By leveraging AI for tasks like these, Ravi not only improves his own productivity but also demonstrates the tangible benefits of AI to his colleagues. This approach helps in fostering a culture of AI adoption within the team.
