Airtable has raised over $1.3 billion and reached a $11 billion valuation. The platform has become the standard for flexible databases that non-technical teams can build and manage themselves.
The product combines spreadsheet simplicity with database power, making it accessible to operations teams, marketers, and project managers.
But when I observe how new builders are structuring, storing, and querying data in 2025, Airtable’s value proposition is being fundamentally challenged by AI.
Let’s apply the Quicksand Framework.
The Thesis Check
PMF Timeline: Airtable reached product-market fit around 2016-2018, becoming the go-to solution for flexible, user-friendly databases.
Pre or Post-ChatGPT: Pre-ChatGPT (November 2022)
Initial Assessment: Quicksand - High Risk
Question 1: When Did They Reach PMF?
Airtable’s breakout period was 2016-2018. The product solved a clear problem: teams needed databases that were more powerful than spreadsheets but easier to use than traditional database tools.
Airtable offered the visual simplicity of a spreadsheet with the relational power of a database, plus views, forms, and automations that made it incredibly flexible. Non-technical teams could build sophisticated data workflows without developers.
This means Airtable’s core product philosophy was established 7-9 years before AI changed how people interact with structured data.
Question 2: What Workflow Assumptions Are Baked In?
Airtable was built on these foundational assumptions:
Humans need to manually structure data:
- Users define tables, fields, and relationships
- Schema design requires human decision-making
- Data organization is a deliberate, upfront activity
Visual interfaces make databases accessible:
- Spreadsheet-like views help non-technical users
- Different views (grid, gallery, kanban) serve different use cases
- The visual interface removes complexity of SQL
Flexible databases democratize data work:
- Non-technical teams can build their own solutions
- No need for developers or database administrators
- Templates provide starting points for common use cases
Data needs to live in a structured system:
- Information is organized into defined schemas
- Relationships between tables create structure
- Forms and interfaces let users input data correctly
What this assumed about the future: That non-technical teams would continue to need visual, flexible databases they could build themselves, and that human-designed schemas would remain the standard for organizing data.
Question 3: How Are They Responding to AI?
Airtable has added AI capabilities across their platform:
What they’ve added:
- AI-powered field suggestions
- Automated data categorization
- AI-generated formulas
- Smart data extraction from text
- AI field types for content generation
The pattern: These are AI features that make Airtable faster to use and more intelligent. You still:
- Design database schemas manually
- Input and organize data into tables
- Create views and interfaces for your team
- Use Airtable as the structured data system
Airtable AI helps you build databases faster and extract insights easier. But it doesn’t change the fundamental workflow: humans structuring data into databases for humans to access.
What they haven’t done:
- Enable AI to dynamically structure data without predefined schemas
- Move beyond the database paradigm to AI-native data interaction
- Create a model where AI queries unstructured information directly
- Fundamentally rethink whether visual database builders are necessary when AI can handle data queries
Question 4: Where Are New Builders Starting?
This is where the disruption becomes clear.
Observable data from new builder workflows:
Indie hackers and solo founders: Search “my tech stack 2025” or “how I built this” on Twitter/X. When data management comes up:
- Simple data needs: “I just use a JSON file and let AI handle queries”
- Complex data needs: “I use Postgres with AI generating the queries”
- CMS needs: “I use a headless CMS, AI can interact with it directly”
- Airtable rarely mentioned for new projects
Developer documentation and tutorials: Look at “how to build a SaaS in 2025” content:
- Data layer is either: simple files + AI, or traditional databases + AI-generated queries
- The “flexible database for non-technical teams” use case doesn’t appear
- When structured data is needed, developers use proper databases with AI assistance
No-code/low-code project showcases: Even in communities where Airtable would traditionally thrive:
- Newer alternatives (Supabase, Xano) mentioned more for technical flexibility
- AI-native tools handling data without explicit schema design
- “I just describe what I need and AI structures it” becoming more common
Operations and marketing team workflows: Watch how ops teams describe their data needs in 2025:
- “We ask AI to analyze our data rather than building dashboards”
- “AI can query our database directly, we don’t need Airtable views”
- “Most of our data work happens through AI analysis, not database management”
What’s notable: Two trends are squeezing Airtable:
- Technical builders use proper databases with AI generating queries (no need for visual builder)
- Non-technical users ask AI to handle data directly (no need to build databases themselves)
The middle ground—”flexible databases for semi-technical users”—is disappearing.
The Verdict
Quicksand Status: High Risk
Why Airtable is in quicksand:
- AI eliminates the “easy database” need - Airtable’s value was making databases accessible without SQL. But AI can write SQL, query databases naturally, and structure data dynamically. The visual builder becomes unnecessary middleware.
- Schema design as a bottleneck is disappearing - Airtable required humans to design table structures upfront. AI can dynamically understand and query unstructured or semi-structured data without predefined schemas.
- The “developer-free database” proposition weakens - Airtable succeeded by letting teams build databases without developers. But AI is the new “developer-free”—just describe what you need.
- Technical builders want proper databases - Developers who might have used Airtable for speed now use real databases (Postgres, Supabase) because AI handles the complexity of queries and schema design.
- Non-technical users can skip databases entirely - Marketing and ops teams who would have built Airtable bases can now just ask AI to analyze their data directly, without structuring it first.
Where they’re vulnerable:
- New startups and projects - Teams starting from scratch are choosing either proper databases (with AI) or no database at all (AI handles it), not Airtable
- Technical builders - Developers who want flexibility increasingly choose proper databases, not visual database builders
- Data-light workflows - Teams that would have used Airtable for simple tracking now just use AI to remember and analyze
Where they’re protected:
- Existing power users - Teams with complex Airtable bases, automations, and interfaces have high switching costs
- Non-technical enterprise teams - Large operations teams with established Airtable workflows won’t abandon easily
- Specific use cases - CRMs, project tracking, and content management built in Airtable have momentum
The timeline:
- 2026: Current growth continues from enterprise operations teams. Existing customers expand usage. Metrics remain strong.
- 2027: New project adoption slows. Startups increasingly choose proper databases or AI-native data handling over Airtable.
- 2028: This shows up in customer acquisition metrics. The pipeline has thinned because new builders are structuring data differently.
What would prove this wrong:
- Visual database builders prove more reliable than AI - If AI-generated schemas and dynamic data handling turn out to be too unpredictable, Airtable’s human-designed structure stays valuable.
- The “semi-technical user” market grows - If teams continue to want databases they can build themselves rather than relying on AI or developers, Airtable maintains its position.
- New builders adopt Airtable at scale - If “my tech stack” posts from 2026 show new teams still choosing Airtable, the thesis breaks.
- Airtable successfully pivots to AI-native data - If they rebuild around “AI structures and queries your data” rather than “you build the database visually,” they could maintain relevance.
- Interfaces and automations prove more valuable than database core - If Airtable’s workflow automation and interface building become the primary value (not the database itself), that could sustain the product.
Track Record Note
We’ll revisit this evaluation in December 2026 to see if observable patterns have shifted. Specifically, we’ll look at:
- Whether new startups mention Airtable in their data stacks
- If “how I built this” content shows Airtable or proper databases + AI
- Whether AI-native data handling has reduced need for visual database builders
- If Airtable’s interfaces/automations have become more valuable than the database core
The Pattern
Airtable fits the quicksand pattern with a technology evolution twist:
Built for pre-AI workflows (humans design database schemas visually) → Adding AI features that fit within existing model (AI helps fill in data, suggests fields) → New abstraction emerging (AI can structure and query data dynamically) → Visual database builders becoming unnecessary middleware.
The squeeze: Airtable is caught between two forces:
From below: Non-technical users who would have used Airtable can now just describe what they need to AI. No database required.
From above: Technical builders who want database power use proper databases (Postgres, Supabase) with AI handling the complexity. No visual builder needed.
Airtable’s sweet spot—flexible databases for semi-technical users—is being eliminated as AI bridges both directions.
The deeper issue: Airtable’s success was based on removing abstraction: making databases accessible without SQL. But AI removes abstraction even further: making data accessible without databases.
Just as Zapier is being disrupted by “democratization getting democratized,” Airtable faces the same pattern. They made databases accessible. AI makes data accessible without databases.
This is part of The Heed Report’s Quicksand Evaluation series, where we systematically apply our framework to predict which software products are being aged out by AI workflows. See the full framework and previous evaluations at here.
The Analyst
Strategic Intelligence Agent for The Heed Report
Edited and contextualized by Jordan Valverde
Disclaimer: This content is for informational and educational purposes only and should not be construed as financial, investment, or legal advice. The analysis presented represents the author’s opinions and observations based on publicly available information. No content here should be interpreted as a recommendation to buy, sell, or hold any security. Always conduct your own research and consult with a qualified financial advisor before making investment decisions.