Local Internet Labs / 2026

What Comes After the Web Page?

The web was built around pages.

AI increasingly operates around entities, records, relationships, retrieval, context and actions.

Local Internet Labs is exploring what that transition could mean for local businesses, places, information and commerce.

Transition model
  1. Web page
  2. Entity
  3. Record
  4. Graph
  5. Context
  6. AI
  7. Agent
  8. Action
The central question

What Should the Local Internet Look Like in the Age of AI?

A place where ambitious ideas can be explored, prototyped, tested and documented before they become commercial products or infrastructure.

01What we’re trying to learn

Six Open Questions

These are research questions, not conclusions. Each one links to the experiment or program where it is being explored.

02Experiments in progress

Experiment Status

All experiments →
EXP 001Prototype

One Business Record

Research area
Business Records
Last updated
2026-09-02
Prototype
OneBusinessRecord.com
Next step
Define a minimum record schema and render the same record across a listing, a profile page and a machine response.
EXP 002Prototype

The Local Graph

Research area
The Local Graph
Last updated
2026-08-14
Prototype
TheLocalGraph.com
Next step
Model a single city as a connected graph and compare retrieval against flat listing data.
EXP 003Architecture

LocalWebMCP

Research area
MCP & Machine Access
Last updated
2026-07-30
Prototype
LocalWebMCP.com
Next step
Draft a tool surface for entity lookup, geographic scoping and record retrieval, then write example agent transcripts against it.
EXP 004Architecture

Business Record Agent

Research area
Agentic Local Internet
Last updated
2026-07-18
Prototype
BusinessRecordAgent.com
Next step
Prototype agent tasks that read a record and answer questions a website cannot.
EXP 005Question

AI Visibility

Research area
AI Discovery
Last updated
2026-06-30
Prototype
BusinessRecordScore.com
Next step
Enumerate the surfaces and test whether a per-surface measure can be observed consistently.
EXP 006Question

Citable Local

Research area
Citable Local Information
Last updated
2026-07-11
Prototype
CitableEngine.com
Next step
Publish identical facts in structured and unstructured form and observe retrieval behaviour.
EXP 007Prototype

Geographic Intelligence

Research area
Geographic Intelligence
Last updated
2026-06-19
Prototype
USAizz.com
Next step
Route entity retrieval through the explicit hierarchy and compare against filtered retrieval.
EXP 008Question

Internet of Record

Research area
Internet of Record
Last updated
2026-05-30
Prototype
InternetOfRecord.com
Next step
Write the architecture down in full and test whether each experiment maps to a component of it.
03The experiment map

Eight Parts of One Architecture

These are not eight unrelated projects. Each experiment tests a different layer of the same proposed architecture.

  1. EXP 001One Business RecordPrototype
  2. EXP 002The Local GraphPrototype
  3. EXP 003LocalWebMCPArchitecture
  4. EXP 004Business Record AgentArchitecture
  5. EXP 005AI VisibilityQuestion
  6. EXP 006Citable LocalQuestion
  7. EXP 007Geographic IntelligencePrototype
  8. EXP 008Internet of RecordQuestion

Do These Experiments Converge Into an AI-Native Local Internet?

04Research areas

What We’re Exploring

05Flagship program

From Listings to Records

What Could 7 Million Self-Placed Listings Become?

These are not 7 million verified Business Records.

The question is what could happen if that information progressively became more structured, claimable, verifiable, connected and machine-readable. None of these problems are solved.

See the research pipeline →
  1. 7M+ self-placed listingsWhat is actually in the data?
  2. NormalizeWhat information can be normalized?
  3. Entity resolutionHow accurately can duplicate entities be resolved?
  4. StructureWhat structure is worth imposing before verification?
  5. ClaimWhat information requires owner verification?
  6. VerifyWhat makes a Business Record authoritative?
  7. EnrichWhat information should AI be allowed to infer?
  8. Business RecordHow should provenance be recorded?
  9. Local GraphHow should conflicting information be handled?
  10. Local surfacesHow should freshness be measured?
  11. Search + AI + agentsWhat should remain human-controlled?
06Testbed

A Real Network to Experiment With

100+
Marketplace nodes
7M+
Self-placed listings

Local Internet Holdings already operates a network of marketplace and local discovery properties. That existing network creates a potential testbed for exploring structured listings, Business Records, entity relationships, geographic discovery, publishing and machine-readable local information.

This describes a potential experimental environment. It does not claim experiments have been run across the network.

07 / The research architecture

Local Entity → Action

The layers each experiment is testing, and the prototypes exploring them.

  1. 01Local entity
  2. 02Business Record
  3. 03Local Graph
  4. 04Citable information
  5. 05Visibility
  6. 06Machine access / MCP
  7. 07AI agent
  8. 08Action
Explore the full architecture →
08Research initiative

Domains as Nodes

Can strategically organized domains become functional endpoints inside a larger information architecture?

Domain → Node → Network → Infrastructure

2,900+ strategic domains is an owner-provided portfolio metric. It does not imply every domain is deployed or in use.

  1. Domain
  2. Specialized purpose
  3. Structured data
  4. Shared entity model
  5. Connected graph
  6. Network node
  7. Human + machine experience
09Prototypes

From Idea to Prototype

Full gallery →
10Build in public

Building the Local Internet in Public

  • Research Notes2026-09-02

    What Comes After the Web Page?

  • Business Records2026-08-21

    Why Every Business May Eventually Need a Business Record

  • Business Records2026-07-24

    From Listings to Records

  • Geography2026-07-09

    Why Geography Matters to AI

  • Citations2026-06-27

    Can Local Information Become Citable Infrastructure?

  • Agents2026-08-07

    The Local Web Was Built for Browsers. What Happens When Agents Arrive?

11Method

Question → Architecture → Prototype → Experiment → Learn → Build → Deploy

01
Question

Start with a fundamental question.

02
Architecture

Design a potential system.

03
Prototype

Make the idea visible.

04
Experiment

Test assumptions.

05
Learn

Document what works and what doesn't.

06
Build

Develop promising concepts further.

07
Deploy

Move mature ideas into the broader Local Internet ecosystem.

12Research principles

How We Think

  1. 01Separate hypothesis from evidence.
  2. 02Prototype before overbuilding.
  3. 03Make architecture visible.
  4. 04Document what doesn't work.
  5. 05Keep humans in control of authoritative business information.
  6. 06Distinguish structured facts from AI inference.
  7. 07Treat geography as context, not just navigation.
  8. 08Build for people and machines.
  9. 09Don't confuse deployment with validation.
  10. 10Ask better questions before claiming answers.
13Handoff

From Experiment to Application

  1. Local Internet Labs
    Research → Experiment → Prototype
  2. Validated concept
    Evidence, not enthusiasm
  3. Local Internet Group
    Strategy → Technology → Services → Implementation
  4. The Local Internet ecosystem
  • Labs asks the questions.
  • Prototypes make the ideas visible.
  • Evidence determines what deserves to move forward.
  • Local Internet Group can help translate mature concepts into applications and services.

Not every Labs experiment becomes a Group offering.

LocalInternetGroup.com ↗
15Thesis
  1. 01

    The page will not disappear, but the record may become more important.

  2. 02

    Local information needs identity.

  3. 03

    Identity needs structure.

  4. 04

    Structure enables relationships.

  5. 05

    Relationships create context.

  6. 06

    Context improves discovery.

Read the full thesis →

The Local Internet Isn’t Finished.

Neither Are the Questions.

We’re exploring how records, graphs, geography, publishing, AI and agents could reshape local information.

One question. One experiment. One prototype at a time.

Enter Local Internet Labs