The Indian Express
Newsroom intelligence
For The Indian Express, we built a live system for the newsroom. It brings the many sources a desk watches into one place and helps editors decide faster what to cover.
PROPOSAL · REF WZZ-PK-2026-10
An AI Journey Proposal for Punjab Kesari
Every newsroom is now asked for more stories, in more formats, at higher quality, with fewer people. Wizzardzz automations take over the routine work that no longer needs a journalist's judgement in 2026, so a leaner team can do far more: hold readers longer with animated explainers, interactive graphics and slideshows, get cited by AI search, and reach readers directly.
A family of titles
News comes inWizzardzz WorkflowsFormatsWhere readers are
About us
Wizzardzz builds AI for newsrooms
We are an editorial AI company, started by people who have run news desks. We build systems that help a newsroom publish more, faster and in its own style, and reach readers wherever they are. Every tool follows the newsroom's rules, and editors decide what goes out.
- Newsroom automation for wires, drafts and print-to-digital
- Animated explainers, video and interactive graphics
- Getting cited by Google AI Overviews and AI assistants
- Reader reach on WhatsApp, Instagram and Facebook
- Archive search and AI consultancy
At a glance
Four transformations this proposal aims to deliver
Hover or tap a card for more
01
Readers who stay
Explainers, interactive graphics and slideshows on every big story.
Readers spend longer on a page that moves and responds, and come back when the news moves again. Every format is built from your own reporting and drops straight into the article.
02
An explainer in minutes, not two days
Fast video edits that are not "AI-generated" video.
Writer, animator, video editor~2 days
Motion Deskminutes
The same animated explainer used to take a three-person team, a writer, an animator and a video editor, about two days. Motion Desk turns it round while the story is still breaking, from your own fonts, photos and data. No generated faces or footage.
03
Cited by AI
Readers increasingly take the AI answer and never click. We make your reporting the source those answers quote and link back to.
दिल्ली में प्रदूषण क्यों बढ़ा?AI answerSource: your story, linked
Answer-first structure, depth on every topic and clean metadata make Punjab Kesari the source Google AI Overviews, ChatGPT and Gemini quote, in Hindi and English.
04
A faster desk, direct reach
Every source on one ranked screen, and stories pushed to readers.
Every wire, rival story and reporter's WhatsApp lands on one ranked screen. Finished stories go straight to WhatsApp, Instagram and Facebook through an approved Meta technology provider.
Our work
We work with
Outlook
WritePod · DraftOne · WIZ
Outlook started with WritePod, our editorial writing workspace, which ran its full term. The desk now uses DraftOne, which turns agency copy and research into publication-ready drafts in Outlook's house style, and WIZ, which assigns tasks on WhatsApp and follows up until they are done.
2026
The internet is reorganising around AI answers
Readers now ask Google, ChatGPT and Gemini, and read the answer. The answer quotes a handful of sources. Being one of them is the new front page.
- 68%US Google searches with no click, Jan–Apr 2026
Most searches end on Google.
Two in three searches now finish without a visit to any website.
- −60%clicks when an AI Overview shows
AI Overviews take the click.
They appear on more than one in five searches. When they do, clicks to websites fall by more than half.
- −33%Google referrals to publishers, one year
Search sends fewer readers.
Publishers expect a further fall of more than 40% over the next three years.
- 9%of AI citations go to news
A few outlets take the citations.
News citations in AI answers go to a small group of publishers. The rest are invisible.
Hindi answers draw on a small pool of Hindi sources. Punjab Kesari can own that pool before others do.
AI citations
Be the source AI cites
AI platforms quote pages that answer clearly, are fresh, name people and places precisely, and come from a publisher that covers the topic in depth. Each of these can be engineered into the newsroom's daily work, without changing how your reporters write.
Answer firstOptimizer
The answer sits in the opening lines, in a form AI can lift. Micro-edits only, accepted or rejected sentence by sentence.
Fresh, and firstCadence
The desk sees the story first and publishes first. Freshness decides many citations.
Depth on the topicTopicVine
A planned cluster of stories that covers every question readers ask, linked together.
Clear structure and metadataMetaGen · DraftOne
Headlines, summaries, entities and structured data generated for every story.
Video and graphicsMotion Desk
AI answers increasingly show video. Explainers give the platforms more of your work to cite.
दिल्ली में GRAP-3 लागू होने पर क्या बंद होता है?
AI Overview
पंपंजाब केसरी
GRAP-3 लागू: दिल्ली-एनसीआर में क्या खुला, क्या बंद
Ccaqm.nic.in
ग्रेडेड रिस्पॉन्स एक्शन प्लान: संशोधित अनुसूची
जादैनिक जागरण
दिल्ली में प्रदूषण पर सख्ती, निर्माण कार्य पर रोक
अअमर उजाला
GRAP के तीसरे चरण में किन वाहनों पर पाबंदी
Illustration of a Google AI Overview · the goal is to hold a source slot on the questions that matter
Tuning
Citation Watch
A weekly tracker of the Hindi and English questions that matter to you. It checks Google AI Overviews, ChatGPT, Perplexity and Gemini, and shows who gets cited: Punjab Kesari, or a rival. The desk sees where it is winning and which stories to strengthen.
Service · video and interactive
Video from every story, without the "AI video" problem
Readers in Hindi, Punjabi and Urdu increasingly get their news as video. Motion Desk turns stories, data and print pages into animated explainers and shorts, at the pace of the news desk.
Rendered from your content and code, not a synthetic AI-generated video
- No generated footage, and no synthetic faces of real people.
- Real video cuts in too: an anchor shot, a reporter's piece to camera or agency footage, edited in alongside the animation.
- Every frame uses your fonts, colours, maps and photo library.
- Every frame can be edited, so the desk signs off exactly what goes out.
- Fast edits: a market move or a court verdict becomes a video while the story is still breaking.
A 30-second Motion Desk explainer, sketched live: Hindi voiceover with subtitles, the Nifty chart, foreign selling, three causes and an anchor cut. Figures as reported in October 2026. Click to pause.
What Motion Desk makes
Hover or tap a card for more
Story explainers
60–90 seconds on what happened and why it matters.
Script from your copy, scenes from your photos and maps, ready in Hindi and English.
Data in motion
Markets, budgets, prices, rainfall and AQI as animated charts.
Sensex, rupee, budgets and rainfall drawn straight from the numbers, and redrawn when they change.
Map stories
Floods, routes, borders, constituencies and the spread of an event.
Districts, routes and constituencies on your own base maps, revealed step by step.
Timelines
Court cases, careers, disputes and how a story got here.
Every date checked against your archive, so readers see how the story built up.
Election graphics
Live tallies, seat cards, swing maps and winner reels as results come in.
Seat cards and tallies refresh as results are declared, ready for web, social and TV.
Article to Shorts
Every major story gets a 9:16 version for YouTube Shorts, Reels and WhatsApp Status.
A 30–60 second vertical cut of each big story, captioned for sound-off viewing.
Breaking-news cards
A branded video card within minutes of a wire alert, before the full story.
Headline, photo and key line on a branded card, out before the full story is filed.
Web slideshows
Swipeable photo and fact slideshows that keep readers on the page.
Photo and fact slides built from the story, each slide its own share card.
Podcast and audio shorts
Captioned audiograms cut from interviews, podcasts and anchor clips.
Waveform, captions and speaker names over the best 30 seconds of a conversation.
Quote and statement cards
Speeches and statements as kinetic captions, attributed and timestamped.
The exact words, the speaker and the time, checked before they go out.
The morning edition, as video
The day's top print stories as a two-minute video brief at 7 am.
The front page as a short video brief, on WhatsApp before readers open the paper.
From the archive
"This day in history" and anniversary films from your own archive.
Anniversary films from decades of your own photos and reports.
Interactive graphics
Calculators and explorers readers operate themselves.
Readers change a number and see what the story means for them. Two examples follow below.
Sports recaps
Cricket and kabaddi recaps built from scorecards, minutes after the match.
Key moments and stats from the scorecard, cut into a recap minutes after the final ball.
Every language
Hindi, Punjabi, Urdu and English from one master.
One master edit, with captions and voice in each language.
How a Motion Desk video is madeEditor checkpoints marked ✓
- Story or dataWire, article, print page or dataset
- ScriptYour editor's script, or one auto-generated from the story✓ Editor
- StoryboardScenes, maps, charts✓ Editor
- Build scenesYour brand templates and assets
- Voice and captionsYour anchors or a licensed voice
- Render16:9 · 9:16 · 1:1 · 4:5✓ Sign-off
- PublishYouTube, Instagram, WhatsApp, web
Traditional team · writer, animator, video editor≈ 2 days
Motion Deskminutes
A managed service
We make it for you
Send a story, a dataset, a print page or your own script. Our team returns finished explainers, shorts, slideshows and interactive graphics in your style, on an agreed daily or weekly volume. Ready to start now.
Examples, made entirely by our video harness
Interactive embeds
Graphics readers operate themselves
Most desks can't build these on deadline. A reader changes a number and sees what the story means for them. They stay longer, share more, and come back when the news moves again. Each one drops into any article with a single line of code.
Demo stories. The numbers in both examples are illustrative and have not been verified. They show how the format works, not reporting.
The story
The RBI raised the repo rate by 0.25 points to 5.50% on 7 October 2026. Floating-rate home loans follow.
What the reader does
Sets their own loan amount, tenure and interest rate, and tries a rate change. The EMI and the extra interest update instantly.
Built for Punjab Kesari
Hindi first, English in one tap. Works on any phone, loads fast, and carries the story link.
Ready for every big number
Fuel prices, budgets, rupee, gold, election swings, monsoon rainfall.
<iframe src="…/embeds/emi/" width="100%"></iframe>
The story
Foreign investors are selling Indian shares while Indian institutions buy. Who is on each side, and where do household savings fit?
What the reader does
Scrubs a timeline to find the day both sides flipped, guesses the one industry foreigners bought, and opens each savings measure.
Built to explain
Every choice rewrites the narration. Listen in Hindi or English. Light or dark. Keyboard and touch.
Honest by design
Different measures keep their own scales, and the sample data is labelled as unverified on screen.
Try them: open the EMI calculator and the Great Money Migration story on the web version of this proposal.
The plan
Built around Hindi, Punjabi and Urdu
Tools that read, sort and route news work in every language from day one. Tools that write copy are tuned with a style sheet for each language and calibrated with your editors. The plan follows that order.
हिन्दीਪੰਜਾਬੀاردوEnglish
Ready in every languageThe live desk. Everything the desk needs to see arrives in one place, scored for your readers, and routed to the right person. Select a source to follow its path.
- Comes in
- Hindi wiresPTI भाषा, UNI वार्ता, IANS
- English wiresPTI, Reuters, AP
- Rival publicationsHindi, Punjabi and Urdu, as they publish
- Reporters' WhatsAppText, photos, voice notes, transcribed
- X postsLeaders, officials, agencies
- InstagramPublic accounts you follow
- EmailPress releases and tips
- Government releasesStates and Centre
one screen- Desk gets
- A score on every storyTwelve editorial measures, tuned to your readers
- A shared priority boardRanked, flagged, timed from arrival
- Workflow trackingWho has each story, and where it stands
- Meta, summary, five key pointsGenerated for every story
- Straight to your CMSOne click, no copy-paste
- Live analyticsWhat the desk picked up, and how fast
Live newsroom analytics
A dashboard of what came in, what was used and how fast, by desk and by person.
Every story tracked
Each story's journey is recorded: who picked it up, edited it and filed it, and when.
Role-based workflow
Reporters, sub-editors, desk heads and admins each see and do what their role allows.
Meta, summary, five key points
SEO and AI-search metadata, a summary and five key points, generated automatically.
Direct CMS integration
Stories go to your CMS in one step. Easy to connect to the system you already use.
Priority board
The day's important stories, ranked and colour-coded by how long they have waited.
Why it fits
Cadence reads, sorts and routes. Language is not a barrier. Your desk can use it within weeks.
English nowHindi tuning in phase 3
An agentic newsroom system that turns a wire story or a brief into a publication-ready draft in house style. Editors approve the research and the structure before a word is written. Select a step to see the screen.
Editorial planning
Editorial planningPlan coverage on evidence, not instinct
AI platforms cite publishers that cover a topic in depth. TopicVine does two things. It builds a topic tree of the subtopics and entities around any subject. And it collects the real questions people ask on Google, scores each one for relevance, and clusters them. The scoring is fixed and repeatable, so the same input gives the same plan. The result is a coverage plan the editor can defend.
Subtopic
Questions of the kind TopicVine collects · scores illustrative
Real questions
Built from what readers search, in Hindi and English.
Scored and filtered
Each question gets a relevance score. Filter to the core.
A linking map
Which story links to which, so authority builds up.
A coverage calendar
What to publish this week, and what is missing.
Consultancy
Archive vectorisation: the first step of any AI journey
Decades of Punjab Kesari reporting are the group's deepest asset. Vectorising turns every story and photo into a searchable "meaning map", so AI can find what the archive knows, not just what it says word for word.
How vectorisation works · an illustration
How we do it for your archive
- Audit what exists: text, scans, photos, rights.
- Digitise and read scanned pages, including legacy Hindi fonts.
- Clean and remove duplicates.
- Split and tag with people, places, dates, bylines and rights.
- Embed text and photos with models that understand Hindi, Punjabi and Urdu.
- Store in a governed vector database, on your own servers if you prefer.
- Connect to the newsroom tools and to licensing.
Search by meaning
"Floods in Punjab in the 1980s" finds the right stories even when they use different words, in any of your languages.
Grounded AI answers
Research and fact-checking tools answer from your own archive, with links to the original stories.
Context on every story
New stories automatically link to the most relevant past coverage, which also builds AI-search authority.
Recommendations from day one
A story can be recommended the minute it is published, by its meaning, before it has any clicks.
Photo search by description
Find archive photos by what they show, not by file name.
An asset to license
A clean, tagged archive is ready to license to AI companies, on your terms.
Generative recommendations
Beyond "more like this", AI can write the recommendation itself, grounded in your archive:
- "Why this matters" context boxes drawn from past coverage
- Personalised reading lists and WhatsApp briefings for each reader's interests
- Story timelines and "the story so far" summaries, built in seconds
- Explainer and archive-video scripts from decades of reporting
What our consultancy coversChoice of models, a cost estimate before any spend, access and rights governance, and readiness to license the archive to AI companies on your terms.
AI harness engineering · print to digital
One print story, broken across pages and editions, stitched together for digital, word for word
A single print story rarely lives in one place. The headline and opening are on page 1, the rest jumps to page 7, the photo sits with the photo desk, the caption in another file, the chart with the graphics team. Transit is built on AI harness engineering: a governed system that finds every piece of one story, stitches them in the right order, checks the result, and lands one complete digital story in the CMS draft folder.
What "AI harness" means
The harness runs AI models inside your rules. It picks from hundreds of skills, checks each result, tries an approved alternative when a step fails, and asks an editor when it is unsure. It never invents a word.
Why not a simple AI workflow
A fixed workflow breaks on the first surprise: a missing jump, an odd file, a photo without web rights. The harness recovers and records every decision.
Failure tolerantIf a step fails, the harness tries an approved alternative, then asks an editor. Nothing is skipped silently.
Never changes the copyEvery word is compared with the print source. Any change blocks delivery.
Editors stay in chargeAutomation on or off by section, bulk approval, and holds that always win.
Partnership
An approved Meta technology provider
- Direct, approved access to the WhatsApp Business Platform, Instagram and Facebook, with no reseller in between.
- News that goes to the reader: breaking-news cards, morning briefings and two-way reader conversations at scale, within Meta's rules.
- A newsroom on WhatsApp: reporters file, readers send tips, and the desk receives both in Cadence.
Meta, WhatsApp, Instagram and Facebook are trademarks of Meta Platforms, Inc.
Built on WhatsApp
The desk's task manager
Task management that lives where the newsroom already talks. Assign it, and stop chasing.
- Assign in plain language. No forms or templates. If a detail is missing, WIZ asks for it.
- Automatic follow-ups. WIZ nudges people when a task is overdue, slipping or stuck, at times you set once.
- Asks for the reason. When work stalls, WIZ checks what is holding it up and passes it on.
- Every status recorded. Each task's journey is logged, from assignment to delivery. No ambiguity, no excuses.
- Status reports any time. Ask WIZ on WhatsApp for a summary of a task, a person or the whole desk.
- Works with your calendar. Connects to Google Workspace and Microsoft 365 and plans around busy schedules.
Everything, by readiness
What works now, and what we tune for you
| Tool | What it does for Punjab Kesari | Status | Phase |
|---|---|---|---|
| Works now | |||
| Motion Desk | Animated explainers, shorts and slideshows made by our team, in your style. | Ready | 1 |
| Interactive embeds | Calculators and explorers readers operate, built for each big story. | Ready | 1 |
| Cadence | The live desk: every source on one screen, scored, tracked and routed. | Ready | 1 |
| WIZ | Assigns desk tasks on WhatsApp, follows up and reports status any time. | Ready | 1 |
| WhatsApp, Instagram, Facebook | Breaking-news cards within seconds, and reader briefings people can reply to. | Ready | 1 |
| DraftOne, English | Publication-ready English drafts from wires and research, in house style. | Ready | 1 |
| TopicVine | Coverage plans built from real reader questions, scored. | Ready | 1 |
| Tuned for Punjab Kesari | |||
| Citation Watch | Weekly tracking of who AI platforms cite on your key questions. | Tuning | 1 |
| Archive vectorisation | The archive made searchable by meaning. Consultancy, then build. | Tuning | 1–2 |
| Feed automation | Weather, AQI, markets, cricket and government releases, drafted to template. | Tuning | 2 |
| Print to digital | Each print story to the CMS draft folder, word for word, with legacy fonts converted. | Tuning | 2 |
| Optimizer, MetaGen | Answer-first micro-edits and AI-search metadata on every story. | Tuning | 2 |
| Fact Checker | Checks claims, including viral WhatsApp forwards, with sources. | Tuning | 2 |
| Language tuning | |||
| DraftOne, Hindi, Punjabi, Urdu | Drafts in each title's style, after a style sheet and a review loop with your editors. | Tuning | 3 |
| Cross-language publishing | Stories adapted between titles and languages, approved by an editor. | Tuning | 3 |
Roadmap
Twelve months, three phases
Phase 1 · Months 0–3
Engage readers, be seen first
- Motion Desk
- Interactive embeds
- Cadence
- Citation Watch
- TopicVine
- WhatsApp, Instagram, Facebook
- WIZ, DraftOne for English
- Archive audit
Phase 2 · Months 3–6
Publish more of what you have
- Print to digital
- Optimizer, MetaGen
- Fact Checker
- Feed automation
- Archive vectorisation
Phase 3 · Months 6–12
Write in every language
- DraftOne in Hindi, Punjabi, Urdu
- Cross-language publishing
- Archive videos and licensing
Next steps
How we start
1
A meeting with your editorial, digital and video heads to agree priorities.
2
A sample Motion Desk explainer and an interactive graphic on stories you choose, and a Cadence demonstration on your own feeds.
3
Scope, cost and timeline in writing.