Skip to content

Why Data Ecosystems Will Determine AI's Winners: Insights from Industry Leaders

Industry leaders reveal why data infrastructure, not AI models, will determine winners. Learn how to build competitive data ecosystems for AI success.

Data-Ecosystem

Why Data Ecosystems Will Determine AI's Winners: Insights from Industry Leaders

Industry leaders reveal why data infrastructure, not AI models, will determine winners. Learn how to build competitive data ecosystems for AI success.

Data-Ecosystem

While the world fixates on AI models, the real competitive advantage lies in the data ecosystems that feed them. At this year's HubSpot for Startups AiSummit in San Francisco, leaders from HubSpot, Amplitude, Frame AI, and Snowflake gathered to break down how data infrastructure, accessibility, and context are becoming the true differentiators in determining who wins in the AI era.

 

Joining the conversation were Karen Ng from HubSpot, Francois Ajenstat from Amplitude, George Davis from Frame AI, Chris Child from Snowflake, and Matt Village from Mindstream as moderator.

Data Is Everything in AI Success

When asked how much of AI success comes down to data, Chris Child from Snowflake didn't mince words: "Ultimately, I think it's really all of it." The panel emphasized that while anyone can plug into ChatGPT APIs, the real competitive advantage comes from bringing deep understanding of your business, customers, and historical patterns to AI systems.

Key insights on data's critical role:

  • Generic AI implementations provide the same experience as competitors
  • True value comes from business context, customer understanding, and interaction patterns
  • Data silos remain a persistent challenge for enterprises
  • Success requires combining data into a unified pool for AI activation

Breaking Down Data Silos Creates Exponential Value

Francois Ajenstat highlighted how connecting disparate data sources creates multiplicative value: "When you're able to connect the dots across all of these various systems, all of a sudden, you're able to spot patterns that wouldn't be possible in a siloed way." The panel noted that adding just one extra data source can multiply impact by 10x at minimum.

The conversation revealed three distinct categories of valuable data:

  • Structured data: Traditional CRM data, contacts, companies, invoice prices
  • Unstructured data: Conversations, emails, images, video, voice calls
  • Streaming/behavioral data: Click streams and real-time interaction patterns

Enterprise Data Provides Unique Competitive Advantage

The panel distinguished between generic, publicly available data and proprietary enterprise data. As Francois explained, enterprise data represents "unique, proprietary, differentiated data versus what's available to everybody else." This includes company-specific taxonomy, semantics, and contextual information that can't be replicated by competitors.

Enterprise data advantages include:

  • Company-specific definitions and terminology
  • Unique cultural context and processes
  • Proprietary interaction patterns
  • Historical business intelligence

Timeliness and Quality Both Matter

George Davis from Frame AI emphasized the importance of data timeliness, describing a three-stage data lifecycle:

  • Early stage: Highly business-relevant data with multiplicative value
  • Secondary stage: Situational importance, activated by specific queries
  • Long-term stage: Audit trails and compliance requirements

The panel agreed that having both sufficient quantity and high quality data matters, but Karen Ng noted that the biggest challenges often involve missing entire data sets due to silos, rather than just poor data quality.

The Foundation-Insights-Action Framework

Chris Child outlined a three-step framework that every organization must follow:

  1. Build data foundation: Understand what data you have, establish governance, ensure cleanliness and trust
  2. Generate insights: Use AI to answer questions faster and iterate more quickly
  3. Enable action: Get insights directly into the hands of decision-makers at the moment they need them

The panel stressed that insights sitting in dashboards that people check once are essentially worthless—data must be embedded in existing workflows to drive behavioral change.

Small Businesses Can Still Win with Limited Data

Addressing concerns about data requirements for smaller companies, Chris Child offered encouragement: "Even if you've got a handful of customers, you most likely have more data than you realize at first." He pointed out that ten closed deals actually represent hundreds of emails, meetings, and calls—providing substantial data for pattern recognition that would be impossible for humans to process manually.

Common Early-Stage Data Mistakes

George Davis warned against overly tight coupling between data models and current product operations. His key insight: "AI is increasing the number of different use cases you can apply the same data to," meaning startups should prepare their data infrastructure for analytic use cases and AI examination much earlier than traditionally recommended.

Francois Ajenstat added that the rapid evolution of AI models—with new capabilities emerging every six weeks—requires product teams to go deeper than traditional product management, combining customer problems with evolving model capabilities.

Revolutionary Impact on Product Development

The panel shared how their companies are integrating AI throughout product development:

Snowflake focuses heavily on engineering productivity, seeing massive boosts in developer output while pushing adoption across resistant veteran developers.

Frame AI uses AI across the entire product development lifecycle—from brainstorming and customer feedback analysis to prototyping and documentation, emphasizing transparency and accountability for outputs.

Amplitude took the most aggressive approach, rebuilding everything "AI-first" with a "burn the boats" mentality, resulting in dramatic speed improvements where tasks that previously took weeks now take days.

HubSpot emphasizes AI fluency and speed as competitive advantages, exemplified by their four-week collaboration with OpenAI to help define the MCP spec for ChatGPT integration.

Data Sharing and Ecosystem Challenges

When addressing barriers to effective data sharing across partners, the panel identified two primary obstacles:

  • Trust issues: Organizations remain nervous about sharing data despite potential mutual benefits
  • Technical challenges: Data often remains locked in silos, making it difficult to extract and share value

Chris Child noted that even suppliers who work together daily often resist sharing data, despite the clear benefits of collaboration.

The Speed Imperative

A recurring theme throughout the discussion was speed as the new competitive moat. As Francois Ajenstat declared: "Speed is the new moat. We have to go execute faster, deliver faster, learn faster than everybody else. If you don't have that speed, you will die."

This speed advantage comes from AI-enabled rapid iteration, faster insight generation, and the ability to embed real-time intelligence directly into business processes.

Key Takeaways for Startup Leaders

The panel's insights reveal several critical strategies for startup success in the AI era:

  • Think beyond model selection: Focus on building comprehensive data ecosystems rather than just choosing the best AI model
  • Break down silos early: Invest in data integration infrastructure before you think you need it
  • Prioritize data foundation: Clean, trusted, accessible data beats large volumes of messy data
  • Embed insights in workflows: Ensure AI-generated insights reach decision-makers at the point of action
  • Prepare for rapid evolution: Build flexible data architectures that can adapt to constantly improving AI capabilities
  • Embrace speed as competitive advantage: Use AI to accelerate every aspect of product development and business operations

The message is clear: while AI models continue to evolve rapidly, the companies that win will be those that build the most comprehensive, accessible, and intelligently structured data ecosystems to feed these powerful tools.

AI Disclaimer: AI helped us summarize the key points from these videos, and our editorial team reviewed everything to make sure it's clear and correct.

Unlock Full Access

Why wait for months to get access to all the videos when you can get it now? Sign up below and not only will you get early access to the full set of video sessions from this year's summit, but you'll have access to the videos from the past two summits as well; which alone is over ten hours of educational content and insights!  Fill out the form to gain instant access to hours of AI video goodness!

    nicholas-holland-agentic-hero

    Building the Future of AI Agents

    Learn HubSpot's 4-layer framework for AI agents and discover how industry leaders are implementing agentic AI to transform business operations today.

    ai-first-company-panel-hero

    Our Top 10 Takeaways from AiSummit 2025

    Discover our top 10 key insights from AiSummit 2025, including how AI agents are becoming co-workers, outcome-based pricing, and organizational transformation.

    ai-transformation-aisummit

    AI Transformation Across Product, GTM, and Internal Ops

    HubSpot CEO Yamini Rangan and Atlassian CEO Mike Cannon-Brookes share candid insights on leading companies through AI disruption at scale.