Six teams, one view per head

The platform is organised the way Tanishq is run, so every head opens the console that belongs to them.

Platform map

Every module on the platform, the engine behind each one, and what it does on this brand.

Aiera Live · Tanishq call desk

Two live voice agents run on this brand: a sales and enquiry agent, and a customer support agent. Both speak English with natural Hinglish switching, close within a 300 second ceiling and end with a WhatsApp or callback offer.

How a call is recorded

Enquiry and lead desk

Every enquiry gets a call attempt inside the hour, a reason code, and a dated next step. Nothing sits in a spreadsheet.

Live enquiry queue

Customer names are masked on this console. Source, category and ageing follow the live enquiry mix.

Funnel on current inputs

The appointment desk

An answered call is worth nothing until a slot is held. The desk holds the slot on the call, confirms it the day before, and writes the outcome back.

Appointment types held on this brand

Rules the desk follows

• A slot is only offered when the calendar shows it open; the agent never promises a time it cannot hold.
• Every appointment carries a named person at the store, not just a store name.
• A no show is called once, the same day, and re-offered a different window.

Service and warranty desk

Job intake, status calls and completion confirmation, so the customer never has to chase the store.

Escalation rules

• Any job past its standard turnaround is escalated to the store manager the same day.
• A second complaint on the same job goes straight to a senior advisor, never back to the agent.
• Warranty decisions are never communicated by the agent until the service centre has confirmed them.

Messaging desk

Not every conversation deserves a call. Where a message clears the volume more cheaply and more politely, the desk sends a message.

Messaging cost avoidance is modelled at the difference between a messaging contact and a voice contact on Indian rates.

Retention and dormancy

19,00,000 identities have gone quiet on this brand. The engine reads the customer book continuously and produces a dated call list per segment.

Segment sizes are modelled from the 528 strong network and the published customer base. Every rate here is an editable input.

Golden Harvest and Encircle book

Golden Harvest instalments and Encircle membership are the two largest customer books on the brand. The desk protects both, call by call.

How the desk behaves on it

• The agent never quotes a discount or a making charge waiver; it states the balance and books the visit.
• A member who asks to stop is flagged and never called again for the same scheme.

Book sizes are modelled from the 528 tanishq stores, india and overseas and comparable programmes. Each one is replaced with the brand's own figure in the first data session.

Loyalty desk

Tier movement is the cheapest growth on this brand: a member about to slip a tier is worth a call, and a member one purchase away from the next tier is worth two.

What the desk does with it

Campaigns and launches

Each campaign arrives at the desk as a script, a call list and a stop rule, not as a brief.

Category focus in the current window

Campaign names describe campaign types, not confirmed Tanishq plans. Call shares are modelled on comparable windows.

Rate and occasion trigger desk

Daily gold and diamond rate movement, plus the wedding and festival calendar per region.

Rules on the trigger

• A fall of more than 1.5% in the daily rate opens a call window on saved bridal enquiries.
• A rise of more than 2% opens a call window on exchange and buy back enquiries instead.
• Wedding date in the record inside 90 days moves that record to the top of the list.

A trigger only ever opens a call window. It never sends an offer, a price or a discount on its own.

Concierge desk and private viewing

High value Tanishq buying happens in a room, not in a queue. The desk books the room, the advisor and the selection before the customer arrives.

Uplift and routing shares are modelled on comparable appointment led retail. They are inputs, not claims.

Collection lookbook and stylist brief

The desk sends a short, priced set before the visit, so the appointment starts from a shortlist instead of a blank counter.

What a brief carries

Marketing studio · Tanishq

Five campaign images and five reels built for Tanishq from its own range, occasions and audience. Every frame is generated on brand and carries the concept, the channel and the call to action it was written for.

Five campaign images

1200 × 750 · feed

India's most trusted jewellery brand

Tanishq

Hero product still — bridal gold and polki necklace with matching jhumkas

Single subject, no clutter. Used as the anchor frame for the wedding season push and cut down for stories.

Product heroBridal families, 25 to 45Instagram feed · Meta paid

Book a store visit

1200 × 750 · feed

Made for wedding season.

Tanishq

Wedding season campaign frame

Human moment around the product, shot to run through the wedding season calendar in every region.

OccasionBridal families, 25 to 45Meta + YouTube masthead

See the collection

1200 × 750 · feed

528 tanishq stores, india and overseas

Tanishq

Store and service frame

Retail proof for local campaigns: the room, the advisor and the process a customer walks into.

Retail experienceBridal families, 25 to 45Google Performance Max · local

Find your nearest store

1200 × 750 · feed

Certified diamonds and stated exchange value

Tanishq

Craft and proof frame

Macro proof frame used wherever the claim needs evidence: process, tolerance, certification, hand work.

Craft proofBridal families, 25 to 45Brand film stills · PR

Read how it is made

1200 × 750 · feed

5 ranges, one ladder

Tanishq

Range and price ladder frame

Flat lay of the ladder from gifting and solitaires to bridal and wedding sets, used in CRM and catalogue.

RangeBridal families, 25 to 45Catalogue · CRM email

Browse the range

Five reels

Each reel ships as a generated 9:16 film with the beat sheet it was cut to.

Campaign windows these creatives feed

Creative concepts are written from published Tanishq range and occasion language. Visuals are generated for this console and are not shot campaign assets.

AI product portfolio · Tanishq

Five AI products designed for Tanishq specifically, each naming the bottleneck it removes, what gets built, and the one number it moves. Scored on business impact against build effort.

Impact and effort are scored by the build team on the brand's own volumes, and are re-scored at the month six review.

Rate-and-occasion trigger desk

Gold rate moves and the regional wedding calendar are the two strongest Tanishq buying signals. This desk turns each into a dated call window on the exact saved enquiries that match.

How it works, end to end

The value, on Tanishq numbers

Moves: Saved enquiries converted per rate window. Every figure above is computed from the inputs on the value model console, so changing an input changes this table.

Quick demo: a rate dip opens a call window

Who uses it, and why it holds

Systems it needs

Rollout on this brand

Guardrails

• Rate is never quoted on the phone; the call states that terms are published in store.
• Consent captured and logged in the first 15 seconds, with an immediate opt-out honoured across every channel.
• A single stop rule across the estate: the same Tanishq customer is never worked twice for the same reason.
• Scored at desk, region and issue level, never against a named advisor.
• Recordings retained for the agreed window, then deleted; transcripts hold no card or ID data.

Impact 78/100 against 31/100 build effort. Horizon Now, current stage Pilot, payback 5 weeks. Re-scored at the month six review.

Scheme and instalment recovery agent

The instalment savings book is the largest repeat-purchase asset Tanishq holds. This agent keeps accounts paying and books the redemption appointment at maturity.

How it works, end to end

The value, on Tanishq numbers

Moves: Scheme accounts kept active. Every figure above is computed from the inputs on the value model console, so changing an input changes this table.

Quick demo: a missed instalment recovered

Who uses it, and why it holds

Systems it needs

Rollout on this brand

Guardrails

• Read-only access to the scheme ledger; the agent cannot alter a balance, a term or a due date.
• No discount, waiver or term change is offered on a call, ever.
• Consent captured and logged in the first 15 seconds, with an immediate opt-out honoured across every channel.
• A single stop rule across the estate: the same Tanishq customer is never worked twice for the same reason.
• Scored at desk, region and issue level, never against a named advisor.

Impact 81/100 against 39/100 build effort. Horizon Now, current stage Prototype, payback 7 weeks. Re-scored at the month six review.

Enquiry-to-appointment agent

Every enquiry that reaches Tanishq gets a call back inside the hour, in the language the customer wrote in, and ends with a slot held in a named store's book.

How it works, end to end

The value, on Tanishq numbers

Moves: Enquiry to appointment rate. Every figure above is computed from the inputs on the value model console, so changing an input changes this table.

Quick demo: an enquiry called back in the hour

Who uses it, and why it holds

Systems it needs

Rollout on this brand

Guardrails

• No price, discount or making-charge negotiation on the call; the agent states published terms only.
• Consent captured and logged in the first 15 seconds, with an immediate opt-out honoured across every channel.
• A single stop rule across the estate: the same Tanishq customer is never worked twice for the same reason.
• Scored at desk, region and issue level, never against a named advisor.
• Recordings retained for the agreed window, then deleted; transcripts hold no card or ID data.

Impact 80/100 against 41/100 build effort. Horizon Now, current stage Pilot, payback 6 weeks. Re-scored at the month six review.

Dormancy and reactivation engine

19,00,000 identities in the Tanishq book have gone quiet. This engine ranks them daily and hands each store a dated call list with a reason on every name.

How it works, end to end

The value, on Tanishq numbers

Moves: Reactivated customers per month. Every figure above is computed from the inputs on the value model console, so changing an input changes this table.

Quick demo: today's reactivation list for one store

Who uses it, and why it holds

Systems it needs

Rollout on this brand

Guardrails

• No price, discount or making-charge negotiation on the call; the agent states published terms only.
• Consent captured and logged in the first 15 seconds, with an immediate opt-out honoured across every channel.
• A single stop rule across the estate: the same Tanishq customer is never worked twice for the same reason.
• Scored at desk, region and issue level, never against a named advisor.
• Recordings retained for the agreed window, then deleted; transcripts hold no card or ID data.

Impact 88/100 against 28/100 build effort. Horizon Now, current stage Pilot, payback 8 weeks. Re-scored at the month six review.

Conversation intelligence for the estate

Every Tanishq call and chat is transcribed, coded to a reason taxonomy and rolled up to region, store and category, so the rising objection is on the leadership pack the week it rises.

How it works, end to end

The value, on Tanishq numbers

Moves: Objections resolved before escalation. Every figure above is computed from the inputs on the value model console, so changing an input changes this table.

Quick demo: last week's rising objection

Who uses it, and why it holds

Systems it needs

Rollout on this brand

Guardrails

• No price, discount or making-charge negotiation on the call; the agent states published terms only.
• Consent captured and logged in the first 15 seconds, with an immediate opt-out honoured across every channel.
• A single stop rule across the estate: the same Tanishq customer is never worked twice for the same reason.
• Scored at desk, region and issue level, never against a named advisor.
• Recordings retained for the agreed window, then deleted; transcripts hold no card or ID data.

Impact 83/100 against 47/100 build effort. Horizon Now, current stage Prototype, payback 10 weeks. Re-scored at the month six review.

Store network

528 Tanishq stores, India and overseas, read region by region, with demand and staffing pressure next to each other.

Store level signals the desk acts on

• Enquiry volume above what the store roster can answer, two weeks running.
• Appointments booked but not honoured, by store and by advisor.
• Requests for stock the store does not hold, so transfer replaces a lost sale.

Stock and availability

The desk stops promising blind. Before an appointment is offered, the enquiry is matched against what the store actually holds.

What happens when the piece is not there

• The desk names a second store in the same city before it names a date.
• A transfer is raised with a dated arrival, and the customer is called when it lands.
• A repeated gap in the same category at the same store is written into the store audit.

Availability percentages are modelled against the category plan. They are replaced by the live stock feed at go live.

Store audit

Display, stock and process compliance scored on the same rubric everywhere, with a dated action against every gap.

Rubric

One audit cycle covers 25% of the points of sale in a quarter, so every point is seen at least once a year. Scores, actions and ageing are modelled on audit cycles of comparable networks.

Conversation intelligence

Every conversation is transcribed, coded and counted, so the brand learns what customers are actually asking, not what the desk remembers.

What comes out of it every week

Coding is at issue level. No individual customer and no individual advisor is profiled or ranked.

Call analysis

Every Tanishq conversation is transcribed, coded and scored: what the caller wanted, what stopped them, what was promised and what happened next. This is the record the desk, the campaign team and the store network all argue from.

A day off the desk, identities masked

What stopped the call, and what the desk did next

What the transcript is mined for

Transcript volumes and scores are modelled on the current script until the brand's own telephony is connected. Caller identity is masked in every screen and every export.

Caller coaching

Every call the Tanishq desk makes is scored on the same six behaviours, and the lowest scoring behaviour becomes next week's drill. Coaching is written at desk, region and issue level, never against a named individual.

The six behaviours, and where the desk stands

Lowest adherence right now is objection handling at 58%. That is the drill below.

Region scorecards

This week's drills, written from lost conversations

How a drill reaches the floor

• Monday: the week's calls are scored and the lowest behaviour is picked by number, not by opinion.
• Tuesday: three good calls and three weak calls on that behaviour are clipped, with identities masked.
• Wednesday: the script line is rewritten and pushed to both the human desk and the voice agent.
• Friday: the same behaviour is re-scored and the movement is reported to Head of retail.

Scores are modelled on the current script and comparable Indian voice desks until the brand's own recordings are connected. No individual advisor is ranked, scored or named in any output.

Quality record

Every conversation is scored on the same five dimensions, at issue and store level, so quality is a number the brand can argue with.

Market sense · web intelligence

Six buckets, read off the open web for Tanishq: competitor moves, category demand, the brand's own online reputation, the price and offer message in market, retail expansion, and the social conversation. Each bucket opens with the analysis — how the brand is actually doing on that front and what to do next — and the sources it was written from sit underneath it as the evidence.

Who is tracked against Tanishq

Category read as: jewellery retail India. Reputation read from mouthshut.com, consumercomplaints.in, trustpilot.com and open consumer forums.

Written off 60 sources for Tanishq. Every claim is backed by the linked sources under each bucket.

How each bucket is built

Results are ranked by how sharply they read against the brand, so complaints and competitive moves surface first. Severity is a keyword read on the headline and summary, not a judgement.

Value model

Move any input. Everything recalculates. Every figure starts from the published Tanishq network and is editable.

Service saving is modelled at ₹42 of handling cost avoided per conversation the desk closes without a store call back.

Data readiness

What is connected on day one, what is modelled today, and what Tanishq needs to hand over for the pilot to go live.

What is needed to go live

• A contactable customer extract with consent flags, in any format.
• Store and advisor calendars, or a single sheet per region to start.
• One session with the Tanishq team to replace every modelled volume with the brand's own.

Required inputs

What the Tanishq team hands over, in whatever format it already exists, for the pilot to run on the brand's own numbers.

Nothing here blocks the console. Every input replaces a modelled figure already on the screen.

Method and limits

How the pilot runs on Tanishq, week by week, and what the system will not do.

What this system will not do

• It will not quote a price, a discount or a making charge outside the published range.
• It will not claim a warranty outcome before the service centre has confirmed it.
• It will not call a customer who has declined or opted out, for any reason.
• It will not score, rank or profile an individual customer or an individual advisor.
• It will not pretend to be human when asked directly.

Commercial terms

One flat retainer for the platform and the desk, usage billed as pass through at cost, and a review at month six.

Governance

• No individual customer is scored, profiled or ranked. Quality sits at store and issue level only.
• Consent and do-not-call flags are respected on every attempt, with a stop rule per segment.
• Every conversation is retained for a stated window and is available to the brand team in full.