It's hard to know what AI can actually do to help your business, but there are some practical applications emerging across industries.
Codingscape product leader, Lauren Mason Carris, and senior AI/ML engineer, Zachary Burkett, got together to talk about cost efficient ways to use AI right now.
They cover everything from why not everyone needs a chatbot to the future applications of AI.
Watch their conversation here
Here’s an overview of what they talked about in the video.
Must-use AI tools for everyone
One of the most valuable AI tools for businesses today is the large language model (LLM), which handles a wide range of tasks – from automating customer support to processing data.
Everyone should be experimenting with or using ChatGPT, Claude, or another LLM at this point.
Zachary also explained how AI has been used behind the scenes in industries like medical imaging and manufacturing for years.
Not everyone needs a custom chatbot
Lauren and Zachary got into the critical question of when businesses should use AI.
While chatbots have become popular, Zachary pointed out that not every business needs a chatbot. For tasks that can be handled through traditional programming, AI may not provide enough value to justify the cost.
AI is most effective when it solves complex, data-driven problems that can’t be addressed through simpler, rule-based systems.
They also talked about how building custom AI models can lead to unnecessary costs if a business’s needs could be met by off-the-shelf solutions.
AI success stories: Duolingo and AI
Some companies are a natural fit for integrating LLMs in their business – e.g. Duolingo. Learning a new language by talking with an AI agent is one of the best customer-experience use cases so far.
Duolingo integrated LLMs into their mobile app experience to help people learn languages faster. Their new feature, Duolingo Roleplay, lets you build real-world conversation skills with characters on their mobile app.
What makes a good chatbot?
Like Duolingo Roleplay, successful chatbots offer more than just general information.
The best chatbots are tailored to specific contextual needs – like retrieval-augmented generation (RAG) systems – that can pull data from a company’s internal documents to provide more informed, accurate responses.
In some cases, chatbots go beyond text by offering interactive options – allowing users to complete tasks or provide feedback directly within the chat interface.
This customization can significantly improve the user experience.
How big of a problem are AI hallucinations?
Zachary and Lauren discussed the issue of AI hallucinations—situations where AI systems generate inaccurate or nonsensical answers.
Although guardrails like well-designed prompts can reduce the frequency of these errors, they remain a challenge.
Zachary noted that AI research is making progress toward better understanding and preventing hallucinations. They’re still a problem, but we might not always have to deal with them like we do today.
Codingscape’s “Bench” – time for AI innovation
One unique aspect of Codingscape is how we use bench time for engineers in between projects.
When team members have downtime between client projects, they work on internal AI initiatives, which fosters both upskilling and innovation.
Lauren and Zacahary highlighted a few AI tools developed during bench time, including:
- Document parsing tools for extracting information from files.
- Custom AI vision models for generating code from sketches.
- Augmented reality (AR) applications that combine machine learning with interactive experiences.
These internal projects let Codingscape to develop cutting-edge AI solutions that improve our own software development workflows and bolster our partners' AI solutions.
Zachary’s engineering journey into AI
Zachary began by sharing his path into AI and machine learning. His early interest started with generative adversarial networks (GANs), which allowed him to experiment with AI-generated images.
Over time, this passion evolved into working with advanced machine learning models, which are now his primary focus at Codingscape.
AI’s future potential
Both Lauren and Zachary expressed optimism about AI’s future.
Zachary emphasized that AI has the potential to revolutionize industries by improving our understanding of the world around us.
From curing diseases to solving complex scientific problems, AI could unlock new possibilities that we haven’t even imagined yet.
Watch the full talk here>
