After Airtableās events, I always like to share my thoughts with my team internally at Claribase. I thought I'd share the recap here as well for folks at the event and for those who missed it. š Welcome The day kicked off with breakfast meet-and-greet. @Howie, Airtableās CEO, opened with insights on the exciting new Airtable AI features and his vision of how itās a game-changer for enterprises. One of my favorite usecase for Airtable AI was Howieās strategic analysis of financial analyst reports. Howie pulls the data in the equity analyst reports into Airtable which reflects the sentiment towards Airtable. He has 10 different questions that he cares like āHow Airtable is represented in the analyst coverage.ā The reports are long and would normally take many hours of research to aggregate the information. Itās such a great example of leveraging AI to help support more strategic decisions. Airtable AI Demos The opening was followed by a show and tell of creative use cases of Airtable AI by Andrew Busse. 1ļøā£The first use case Andrew showed was a podcast/advertiser matching demo for a company that creates podcasts. Each podcast this company makes has a certain amount of advertisement spots in each episode and the company also maintains a list of advertisers they work with. Within this context, Airtable AI was leveraged to match advertising spots in a podcast episode with a relevant advertisement through with the help of various parameters like podcast genre, the podcastās script, the reach of the ad slots, and the amount of ad slots. Additionally, the AI tool was used to script the ad introductions, transforming them into ready-to-use scripts for seamless podcast integration. It was an interesting demonstration of the strategic capabilities underlying Airtable AI and it showcased the AIās ability to understand the context of the data in a record to match it accordingly. 2ļøā£The second use case demonstrated Airtable AIās translation capabilities. The team showed how AI could be used to translate product descriptions for international sites at scale. Even on a first pass, Airtable AI had 92% accurate translationsāa huge time saver. This was a great example of how weāll likely see more āmanualā (traditionally human) tasks supported by AI. Moreover, the team demonstrated how the same AI could flag any translations it knew to be weak or potentially not fully correct, and tag the records, which then went to a review page inside interfaces for human review. Within this context, the ability to get the AI to review its own work and introduce a human review component makes this a unique and interesting use case of Airtable AI. 3ļøā£The third use case explored how Airtable AI can be used within the natural flow of a business' operations to speed up preparing summaries and standardizing reporting. In this example, the team demonstrated how the AI could write summaries of user feedback, take information from surveys - gather sentiments, prepare press releases and documentation, and overall save the management team time documenting, summarizing, and recording data. Additionally, when a user gives feedback, AI could also be used to generate a response email and provide support. Overall, these use case demos were a great showcase of the depth and breadth of the types of use cases weāll likely encounter in the coming months as we start to integrate Airtable AI into more projects. Airtable AI Presentations After the kickoff, the event continued with two presentations: Airtable AI & Security: In the first presentation, Chelsea Tenaka highlighted the meticulous measures Airtable takes to ensure data security and privacy in its AI features. She detailed the logistics of how information and data used in AI prompts and datasets are securely managed to prevent their use in model training. This is very important for getting tools approved internally. Prompting 101: Mikayla LaRosa's presentation on "Prompting 101" emphasized treating interactions with Airtable AI as conversational, defining the AI's role clearly, and providing explicit instructions and constraints to achieve optimized outputs. Airtable AI Build Post-lunch, we transitioned into an AI Workshop, where the participants had a chance to team up, and test and build new tools with Airtable AI. We create a duplicate checker within Airtable's automation & AI features. We really liked how the room built demos that leverage AI and also included steps for human intervention to mimic real life workflows! Shoutouts and Thank You's As always, thank you to @dariot , Kelly, and Alisson - the amazing Airtable team that puts these events on. Thank you all for the great talks, coffees, and connections. I always love traveling to SF - but this trip was really special. The overall atmosphere at Airtable HQ was wonderfully collaborative, and felt like the grassroots community event that Iāve always loved the Airtable community for!
... View more
