Dec 19, 2017 · 21m · mad
Superpowered Conversations // George Davis, Frame.ai (FirstMark's Data Driven)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
At FirstMark's DataDrivenNYC event, Frame.ai CEO George Davis explains how modern team chat represents a major data transformation, detailing Frame.ai's assistive AI platform designed to parse operational conversations and supercharge B2B customer success.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
Matt as informed peer 0.0 Guest teaching 0.0 Guest disagreement 0.0 Matt pushing back 0.0
speaking balance: gold is Matt, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 18:45 Acknowledging hard limitations of AI conversation handling
Davis frankly points out the difficulty of the problem, noting that anything hard for humans remains inherently difficult for automated AI systems.
Hardest push from Matt ▶ 21:00 Moderator wrap-upThe event host/moderator steps in to end the session and transition to offline post-talk discussions.
Biggest teaching moment ▶ 19:34 Differentiating structured summaries from standard NLPDavis educates the audience on Frame.ai's specialized approach to machine-readable structured summaries compared to traditional NLP summarization techniques.
Matt holds his own ▶ 15:20 Deep dive into semantic frames and ML architectureDavis demonstrates domain expertise by explaining how semantic frames serve as the junction point between unstructured chat and structured transactional tools.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Matt as informed peer | Guest teaching | Guest disagreement | Matt pushing back | Why |
|---|---|---|---|---|---|---|
| Conversation is Eating the Enterprise | 0 | 0 | 0 | 0 | This segment is a solo presentation monologue by George Davis introducing Frame.ai and asking the audience about Slack usage. Because the host does not participate, all host scores are 0. | |
| Moving from Email Threads to Dedicated Channels | 0 | 0 | 0 | 0 | Davis delivers a solo presentation segment comparing email threads to channel-based team chat workflows. The host does not engage, requiring host metrics to remain 0. | |
| Defining the Enterprise Graph | 0 | 0 | 0 | 0 | Davis defines the enterprise graph concept during his solo presentation deck. Host activity is entirely absent in this monologue segment. | |
| Frame.ai's Virtuous Cycle Architecture | 0 | 0 | 0 | 0 | Davis details Frame.ai's virtuous cycle architecture and B2B customer success use cases in monologue form. With no host dialogue, host metrics are 0. | |
| Operational Challenges in Team Chat Workflows | 0 | 0 | 0 | 0 | Davis explains operational friction points like attention management and data synchronization in enterprise chat. Host scores remain 0 due to the presentation format. | |
| Machine Learning Challenges in Enterprise Chat Analysis | 0 | 0 | 0 | 0 | Davis covers machine learning challenges in semantic extraction and entity relationships during his monologue. Without host interaction, host-side scores are 0. | |
| Audience Q&A Session with Matt Turck | 0 | 0 | 0 | 0 | Audience members ask technical Q&A questions which Davis answers collaboratively. The host only makes brief administrative wrap-up remarks, leaving interaction scores neutral and low. |
Statements from this episode (14)
Davis: Team chat is the biggest current enterprise data opportunity
“I think it's the biggest data opportunity and transformation that's happening right now because it is accumulating a representation of all of the business activities in your firms.”
Davis: Enterprise messaging is shifting from email threads to channels
“What's a little bit newer is popularizing the move from email threads to channels.”
Davis: Integrating systems into chat channels transforms team chat into a platform
“The last thing, and the thing that makes Team Chat a platform, and the thing that's probably changing fastest, is the ability to integrate and put other data and systems in the same room as your conversation.”
Davis: Direct chat integrations eliminate miscommunication and share identical context
“And if you can actually take that system and let it speak for itself, you end a telephone game, ah, and you're able to give everybody exactly the same context.”
Davis: Team chat reveals actual work structure better than org charts
“What's happening is your team chat is collecting a representation of the business activities of your company. And importantly, this is the real structure of the work at your company, not what the org chart or the seeding chart makes you think the structure of …”
Davis: Chatbots are best suited for repetitive tasks lacking user training
“Bots are a very specific interface choice, ah, good in situations where you have repetitive work that doesn't, you can't really afford to do training on, so you need to have a very natural interface.”
Davis: B2B customer success is hungriest for conversational AI assistance
“So when we began talking to people about this blueprint and learning about how they used Team Chat, we looked at a lot of different use cases, and we ultimately found that the place people were hungriest for this kind of assistance was not something we expecte…”
Davis: B2B companies are increasingly moving customer success into Slack
“There is, ah, increasing demand for people to move those conversations into Team Chat, and we see companies inviting their customers into their Slack, opening up shared channels, ah, installing pipelines in between them.”
Davis: Transactional tools like Salesforce and Jira aren't going away
“The transactional workflows that we have in other systems, Salesforce, Jira, Asana, those aren't going away, and a lot of mistakes get made when people discuss something and they consider the work done on email or Slack, and it doesn't get back into those othe…”
Davis: Time to insight is the key metric for enterprise AI
“So the important metric is time to insight. How quickly can we actually be doing things that reflect a real understanding of what's happening in your company?”
Davis: Conversation summaries link unstructured enterprise chat with transactional systems
“Summaries of what's happening in a conversation right now are the interface point between a freeform conversation and a transactional System that needs to talk about that conversation later on.”
Davis: NLP tasks hard for humans remain hard for AI
“Basically anything that is really hard for you to do as a human is still gonna be pretty hard for us.”
Davis: Frame.ai is re-implementing Google's Sling framework
“We're kind of, we're fiddling with and re-implementing the Sling framework that came out of Google a couple of weeks ago we think there's a lot of opportunities there in terms of creative uses of neural networks to represent semantic frames at different timesc…”