Work, Reimagined with AI.

Social Impact · Reuse is the new renewable

Don’t build it
all yourself.

AI made writing software cheap. It did not make running it cheap. If every company prompts its own copy of the same features, the world pays for it in data centers, power, water and billions of dollars of duplicated work.

Herd is built the other way around: tested components, declared in JSON manifests, shared by every customer in the OaaS. Build once. Reuse everywhere.

See the components Earn the badge
Herd Social Impact badge: Built on Reuse Awarded to organizations that run on reusable Herd AI components inside the OaaS instead of rebuilding software from scratch. SOCIAL IMPACT HERD AI · IN THE OaaS Built on Reuse CERTIFIED · 2026

The risk

If everyone builds everything, we’ll need data centers everywhere.

Every rebuilt feature brings its own repository, test environments, AI calls and a maintenance bill that never ends, all running on hardware that has to live somewhere. The numbers are already moving the wrong way.

~945 TWh

Data-center electricity use by 2030, more than double 2024 and slightly more than Japan uses today.1

$660–690B

Planned 2026 infrastructure spend by the five largest US cloud and AI companies, nearly double 2025.2

11.8%

Share of all US electricity that data centers could use by 2030.3

800B liters

Water consumed in 2023 to generate the electricity US data centers used, on top of 66 billion liters for cooling.4

+81%

Rise in duplicated code blocks since 2023 as AI writes more of the world’s code, while refactoring and reuse fall.5

90%

Enterprise software engineers expected to use AI code assistants by 2028, up from under 14% in early 2024.6

The rebuild trap

Same features. Thousands of copies.

Charts, a contact viewer, document generation, a graph canvas, a meeting assistant. Nearly every company needs them. With AI, it’s tempting for each one to generate its own.

We already know what reuse is worth. Without shared open source code, Harvard researchers estimate companies would spend 3.5 times more on software, or $8.8 trillion to recreate what they reuse today.7 AI shouldn’t undo that.

1,000

101001,00010,000

Duplicated build spend

$0.56B–$0.77B

Upkeep every year after

$112M–$154M

Engineering team-weeks

144K–198K

Built in the OaaS

Once

Herd model estimate, not measured customer savings: five components sized at 144–198 team-weeks ($0.56M–$0.77M) to build in-house at about $3,900 per team-week, with upkeep at 20% of build cost a year. Excludes compute, hosting and license fees.

The Herd approach

Build once. Declare it. Share it.

01

Build once

Each Herd component is built, tested and hardened one time, by the team that owns it, not regenerated from a prompt by every customer.

02

Declare it in a manifest

A JSON manifest states what the component does, what it needs and what events it emits. Hosts connect to that contract, so nobody rewrites it to fit.

03

Share it in the OaaS

Every customer runs the same components on Herd: Work. Weekly enhancements reach everyone at once, instead of thousands of teams patching their own copies.

The components

Five you never have to build.

A sample from the Herd library. Each one ships with its manifest. Open any of them to see the real contract.

Herd Charts showing a revenue trend with flagged anomalies

Herd Charts

HerdCharting · v1.0.0

AI-recommended charts, dashboards and anomaly detection.

  • 13 chart types, from heatmaps to geographic maps and network graphs
  • Anomaly detection with IQR and z-score methods
  • Exports to PNG, JPEG, SVG, PDF, CSV and JSON
View manifest excerpt
{
  "component": "HerdCharting",
  "version": "1.0.0",
  "description": "A full-featured data visualization and charting component with AI-powered recommendations, anomaly detection, geo maps, theming, dashboards, and multi-format export.",
  "chartTypes": {
    "total": 13,
    "types": [
      "bar",
      "bar3d",
      "line",
      "pie",
      "scatter",
      "histogram",
      "boxplot",
      "bubble",
      "treemap",
      "heatmap",
      "wordcloud",
      "geo",
      "graph"
    ]
  },
  "anomalyDetection": {
    "methods": [
      "iqr",
      "zscore",
      "spike",
      "levelShift",
      "rateOfChange",
      "run",
      "volatility",
      "rarity",
      "geoOutlier"
    ]
  },
  "exportFormats": [
    "png",
    "jpeg",
    "svg",
    "pdf",
    "csv",
    "json"
  ],
  "persistence": {
    "tables": [
      "herd_saved_charts",
      "herd_saved_dashboards",
      "herd_user_settings"
    ]
  }
}
Herd Contact and Accounts viewer for a demo account, Northwind Aerospace

Herd Contact & Accounts

HerdAccContactViewer · v1.0.0

One viewer for people and companies, with what to do next.

  • Ranked Next Best Actions, health score and lifetime value
  • Pinning, review cards and full edit mode in one tabless view
  • Every interaction emits a typed event to the host
View manifest excerpt
{
  "component": {
    "name": "HerdAccContactViewer",
    "version": "1.0.0",
    "description": "A unified, production-grade viewer component for displaying people (contacts) and companies (accounts) with all their related information in a single, tabless interface with expandable sections, ranked Next Best Actions, health scoring, LTV tracking, pinning, review cards, and full edit mode.",
    "exportName": "HerdAccContactViewer"
  },
  "props": {
    "entityType": {
      "type": "'contact' | 'account'",
      "required": true
    },
    "entityId": {
      "type": "string (UUID)",
      "required": true
    },
    "onEvent": {
      "type": "(event: ViewerEvent) => void",
      "required": false
    },
    "modal": {
      "type": "boolean",
      "required": false
    },
    "editable": {
      "type": "boolean",
      "required": false
    }
  },
  "capabilities": [
    "identityRail",
    "healthScore",
    "ltvIndicator",
    "nextBestActions",
    "enrichedData",
    "notes",
    "meetings",
    "emails",
    "messages",
    "meetingRooms",
    "opportunities",
    "documents",
    "…"
  ],
  "events": {
    "description": "Every user interaction emits a typed event through onEvent callback and is persisted to the activity_log database table"
  }
}
Herd Docs Integration Map showing readiness checks for a component install

Herd Docs

HerdDocs · v2.0.0

Presentations, documents, spreadsheets and text from one brief.

  • Staged pipeline from brief to research, outline, QA gate and render
  • 26 lifecycle events, from generation to editing
  • Integration Map checks data, storage and AI before go-live
View manifest excerpt
{
  "component": "HerdDocs",
  "version": "2.0.0",
  "kind": "composite",
  "intent": "Self-contained AI document generation engine with multi-format output, media library, and WYSIWYG editing.",
  "capabilities": [
    "ai-generation",
    "multi-format-output",
    "media-library",
    "wysiwyg-editing",
    "slide-agent",
    "input-sets",
    "brand-theming",
    "event-bus",
    "pipeline-caching"
  ],
  "outputFormats": [
    "presentation",
    "document",
    "spreadsheet",
    "text"
  ],
  "pipelineStages": {
    "presentation": [
      "brief",
      "research",
      "outline",
      "content",
      "assembly",
      "qa",
      "render"
    ]
  },
  "events": [
    "generation:start",
    "generation:complete",
    "generation:error",
    "generation:progress",
    "slide:edit",
    "slide:apply",
    "slide:navigate",
    "media:add",
    "… 26 total"
  ],
  "integration": {
    "minimalMount": "<HerdDocs />"
  }
}
Herd Graph mind map with branches and topics

Herd Graph

HerdGraph · v1.0.0

Interactive org charts, mind maps and state machines.

  • 6 automatic layouts and 6 node shapes
  • 37 optional event callbacks for full host control
  • Presentation mode, filtering and export built in
View manifest excerpt
{
  "component": "HerdGraph",
  "version": "1.0.0",
  "description": "A fully interactive SVG-based graph visualization component for React with support for nodes, edges, nested children, layouts, theming, filtering, export, and programmatic control.",
  "props": {
    "data": {
      "type": "GraphData",
      "required": true
    },
    "config": {
      "type": "Partial<GraphConfig>",
      "required": false
    },
    "className": {
      "type": "string",
      "required": false
    },
    "eventHandlers": {
      "type": "GraphEventHandlers",
      "required": false
    }
  },
  "enumerations": {
    "NodeShape": [
      "circle",
      "rectangle",
      "rounded-rect",
      "diamond",
      "hexagon",
      "ellipse"
    ],
    "EdgeAnimationType": [
      "none",
      "dots",
      "pulse",
      "circles",
      "dash-march",
      "glow-pulse"
    ]
  },
  "features": {
    "layouts": [
      "force",
      "hierarchical",
      "circular",
      "grid",
      "radial",
      "custom"
    ]
  },
  "events": {
    "description": "37 event callbacks, all optional, passed as props on the component."
  }
}
Herd Meet real-time meeting score with three progress rings

Herd Meet

HerdMeeting · Teams side panel

A live meeting coach inside Microsoft Teams.

  • Prepared briefings keyed to each meeting
  • Live score across people, goals and follow-ups
  • Real-time meeting cost and participant intelligence
View manifest excerpt
{
  "component": "HerdMeeting",
  "source": "FEATURES.md",
  "surface": "Microsoft Teams meeting side panel",
  "features": [
    "meeting-briefings",
    "live-meeting-score",
    "real-time-cost-tracker",
    "participant-tracking",
    "contact-research-cards",
    "goals-checklist",
    "follow-ups",
    "real-time-sync",
    "demo-standalone-mode"
  ],
  "scoreRings": {
    "people": "participants joined / expected",
    "goals": "talking points checked off",
    "followUps": "follow-ups captured / goals"
  }
}

Readiness before release

Manifests make reuse safe.

Because every component declares what it needs, Herd checks the install before it goes live. In the Docs Integration Map, a missing table fails the check, so the problem is fixed once in setup, not discovered in production.

  • Declared capabilities, props and events
  • Live checks on data, storage and AI connections
  • One upgrade path for every customer
Herd Social Impact badge: Built on Reuse Awarded to organizations that run on reusable Herd AI components inside the OaaS instead of rebuilding software from scratch. SOCIAL IMPACT HERD AI · IN THE OaaS Built on Reuse CERTIFIED · 2026

The badge

Built on Reuse.
Certified in the OaaS.

Herd customers who choose shared components over rebuilding earn the Herd Social Impact badge. Put it on your site, your proposals and your sustainability report. It says you didn’t build it all yourself, on purpose.

  1. 1

    Live in the OaaS

    You run on Herd: Work in production, on the shared platform.

  2. 2

    Reuse first

    At least three Herd components are in production in place of software you would otherwise build.

  3. 3

    Declared, not duplicated

    Each component is installed from its manifest, with no parallel rebuild of the same capability.

  4. 4

    Renewed every year

    Herd reviews your components in use and the rebuilds you avoided, and reissues the badge for the new year.

Badge embed code

Download PNG
<a href="https://getherd.ai/social-impact/" title="Herd Social Impact: Built on Reuse">
  <img src="https://jeycypvmeugtxrdaqapd.supabase.co/storage/v1/object/public/media/2d663a4c-037e-4384-bb98-0412262e6839/1792084292710-b880278c-herd-social-impact-badge-1040.png"
       alt="Herd Social Impact badge: Built on Reuse, certified in the OaaS"
       width="140" height="140">
</a>

The reuse pledge

Reuse before you build.

Look first

Before prompting a new feature, check whether a tested component already does the job.

Declare it

Anything new gets a manifest, so the next team can reuse it instead of rebuilding it.

Share the upgrade

Fix it once and ship it to everyone, rather than patching a thousand copies.

Join the OaaS Take the tour

Sources

  1. IEA, Energy and AI: Energy demand from AI
  2. Futurum Group, AI Capex 2026: The $690B Infrastructure Sprint
  3. Lawrence Berkeley National Laboratory, US Data Center Energy & Water Modeling
  4. PLOS Water, Four water insecurity concerns about datacenters (citing LBNL 2024)
  5. GitClear, The Maintainability Gap: AI Code Quality in 2026
  6. Gartner, Top Strategic Trends in Software Engineering for 2025 and Beyond
  7. Harvard Business School, The Value of Open Source Software