137 apps and none of them talk to each other

The average company maintains roughly 137 applications across its business. Knowledge workers use about 8 different SaaS apps daily and toggle between them 1,200 times per day. Each toggle is a context switch. Each switch has a cognitive recovery cost. The aggregate: an estimated $450 billion in lost productivity across the global economy annually (Atlassian).

But the performance cost of switching is only half the problem. The larger issue is that each app creates its own information silo. The client context is in the CRM. The project plan is in the project management tool. The design decisions are in Figma. The technical constraints are in GitHub. The actual agreement about what to build was reached in a Slack thread. And no single search, no single view, no single interface lets anyone see the complete picture.

Why this is the natural state

App sprawl is a bottom-up phenomenon. Each team adopts the tool that's best for their specific work, and these decisions are individually rational. The CRM is better for pipelines than a wiki. GitHub is better for code than a shared drive. Figma is better for design than a document editor. Nobody made a bad choice. The fragmentation is a side effect of everyone making good choices independently.

Top-down consolidation ("let's all use Notion for everything") fails because the specialised tools are better at their specific jobs, and forcing people off them degrades their productivity. The engineer who's told to document their code decisions in Confluence instead of in GitHub PR descriptions will either resist or comply resentfully, and the documentation quality will suffer either way.

What connections change

The approach that works is a connective layer that spans the existing tools without replacing any of them.

Fabric connects to Google Drive, Slack, GitHub, Gmail, Notion, Figma, HubSpot, Linear, Airtable, Salesforce, Asana, Dropbox, and dozens more. Each tool's content becomes searchable from one place by meaning.

Through MCP, this unified knowledge layer also becomes the context for any AI agent. An AI that can search across all your tools understands your company's full picture rather than the fragment visible from any single app.

Each team keeps the tool they prefer. The search spans everything. The silos become transparent without anyone having to change how they work.

Frequently asked questions

Should we try to reduce our app count? Reducing unnecessary tools is sensible housekeeping, but the goal shouldn't be minimising the app count. The goal should be ensuring that the knowledge in each app is findable from one place. Ten specialised tools connected to a unified search layer is better than one general-purpose tool that nobody likes.

How does this handle different naming conventions across tools? Semantic search finds content by meaning rather than exact keywords. "The Q3 client presentation" finds the document whether it's titled "Q3 Deck," "Client Update September," or "Quarterly Review Slides." You don't need consistent naming conventions when the search understands concepts.

What about security across connected tools? Each connection respects the source tool's permissions. Content that's restricted in the source remains restricted in search results. You can configure each connection to include or exclude specific repositories, channels, or folders.

Related reading: Information silos are the default, The cost of scattered knowledge, Too many tools, How to break down information silos. Related pages: Connections, Marketplace connections, One search.