Tristar Leasing: One Car Leasing CRM for 380+ Leased Cars, Linked to Tally and Powered by AI
A corporate car leasing company was running leads, quotes, purchase orders, fleet costs and accounts in separate places. We built one CRM that carries every lease from first enquiry to end of lease, pulls the numbers from TallyPrime, and uses AI to draft quotes and read repair invoices.
- Client
- Tristar Leasing Pvt. Ltd.
- Location
- Mumbai, serving 7+ cities
- Industry
- Vehicle leasing / fleet
- Built
- Web CRM, Tally link, AI assistants
One system for the whole life of every lease
Tristar Leasing now runs its whole leasing business from one CRM. A lead becomes a quote, the quote goes through approval, a purchase order is raised, the car is delivered, and from that day the car’s rentals, repairs, insurance and costs are tracked against it until the lease ends. Finance figures come straight from TallyPrime, and AI does the slow reading-and-typing work: drafting quotes from a dealer proforma, reading repair invoices, and checking the data for mistakes.
About Tristar Leasing
Tristar Leasing is a Mumbai-based company that leases cars to corporates and to their employees across India — Mumbai, Bangalore, Delhi NCR, Pune, Hyderabad and more. Customers pay one fixed monthly rental that covers the car, maintenance and insurance, with no down payment, and employees can take a car as part of their salary package. The company has strategic backing from Mitsubishi Corporation, Japan.
Leasing looks simple from outside. Inside, every single car is a long project: a quote built on GST, interest and residual value; a purchase order with a dealer; KYC and an agreement; delivery; then three to five years of rentals, servicing, tyres, accidents, insurance renewals and finally a buy-out, extension or resale.
One lease, five different places
Before the CRM, that life story of a car was scattered. Leads sat with individual sales people. Quotes were built in Excel sheets that each person kept a little differently. Accounts lived in TallyPrime on an office PC. Repair bills arrived as scanned PDFs and photos that someone had to type in line by line. Nobody could answer a simple question like “what has this car actually cost us so far?” without a day of digging.
Quotes were slow and risky
Pricing depended on spreadsheets, and a wrong cell could mean a loss-making lease.
No single view of a car
Rentals, repairs, insurance and Tally costs were never in one place.
Tally was a black box
The numbers were correct, but only inside Tally, on one machine.
Manual typing everywhere
Repair invoices, cost entries and fleet lists were keyed in by hand.
No approvals trail
Who approved a quote, a lost lead or a price change was not recorded.
From first enquiry to end of lease — one tracked path
Every lead moves along a set path — New, Contacted, Qualified, Quoted, Purchase Order, Agreement, Delivered — and each stage has a time limit, so nothing sits forgotten. The later steps move on their own: raising a PO moves the lead, and delivering the car adds it to the fleet.
- Step 1EnquiryWebsite enquiries land in the CRM automatically
- Step 2QualifiedOwner, priority, next action and reminders
- Step 3QuotedApproved quote from the car catalogue
- Step 4Purchase OrderUn-allotted → Allotted → Disbursed
- Step 5DeliveredCar joins the live fleet automatically
- Step 6Fleet & P&LRentals, costs and end-of-lease per car
A car leasing CRM around the real workflow
We did not start from an off-the-shelf CRM and bend it. We mapped how a lease actually moves through Tristar — sales, operations, accounts, after-sales and RTO — and built the custom CRM around those steps. Every module below is in daily use.
Leads and sales pipeline
Enquiries from the Tristar website land in the CRM automatically.
- Sales sees them on a Kanban board, can assign and bulk-reassign, log calls and set the next action, and gets reminder pop-ups when a follow-up is due.
- Marking a lead as lost needs a manager’s approval, so good leads are not quietly dropped.
Quotes with approval built in
Quotes are generated from a car catalogue of 26 makes, 241 models and 1,614 variants, so make, model and variant are always picked, never typed.
- Every quote goes into an approval queue.
- A rejected quote comes back with clear correction codes and can be revised in one step; an approved quote becomes a ready-to-send PDF.
- For complex deals, the team can open a lead’s quote as a spreadsheet pre-filled from Tristar’s own GST template — 5%, 18% and 40% GST sheets plus a lease amortisation sheet — right inside the browser.
Purchase orders and delivery
A purchase order records the customer and end user, car, colour, fuel, dealer, prices, monthly rate and delivery date.
- Its status moves from Un-allotted to Allotted, Disbursed and Delivered.
- Raising the PO moves the lead forward on its own, and marking it Delivered adds the car and driver to the live fleet.
- Changes to the PO date are logged with a reason, and remarks are kept as a history.
Fleet and post-delivery management
After delivery, the car joins the live fleet register.
- Operations can track rental payments, end-of-lease decisions (exit, fixed extension or holdover), early terminations, hypothecation releases, dealer reconciliation and cess recovery, and generate NOC and NDC letters as PDFs.
- Alerts flag what needs attention this week, and a turnover view projects the financial year.
Cost tracking and used-car P&L
Every repair, tyre, battery, insurance and service cost is tied to the car it belongs to — about 12,000 cost entries so far.
- Costs can be imported in bulk with a preview and a one-click roll-back if a file was wrong.
- The system then shows running cost per kilometre by make, pricing bands from past quotes, and the real profit on each used car sold at the end of a lease.
No lead forgotten, no lead quietly dropped
The lead module is built around one rule from Tristar’s sales head: every lead has an owner, a next step and a deadline.
Seven stages with time limits
New, Contacted, Qualified, Quoted, Purchase Order, Agreement and Delivered — plus Lost and Inactive.
- A lead can only move along the allowed path
- Time limits per stage: New 2 days, Contacted 5, Qualified 10, Quoted 7
- List, Kanban (drag and drop, bulk move), Won and Team performance views
Priority and source on every lead
Hot, Warm, Cold, General or Confirmed — and where it came from.
- Sources: walk-in, referral, website, social media, cold call, exhibition
- Calls, emails, meetings and notes logged on the lead’s timeline
- Documents (PDF, Word) attached to the lead
Next action and daily reminders
Each lead carries a next-action date and note.
- A pop-up each day lists the leads due today — read, dismiss or snooze 30 minutes
- A notification centre and a badge in the menu show the count
Lost leads need two approvals
Marking a lead lost needs one of eight reasons, then approval.
- Reasons: competitor won, price too high, timing off, no decision authority, unreachable, budget cut, product mismatch, other
- Both the sales admin and the master admin must approve; a reviewer can reject it back
- A lost lead can be revived later
Assignment by role
Sales admins assign and bulk-reassign leads within their own team; sales people see only their own leads.
- Operations can see qualified and quoted leads to prepare in time
Website enquiries, straight in
Enquiries from tristarleasing.in arrive through a secure, token-protected link.
- The company is matched or created automatically
- A new lead is opened with source “Website” and the message attached
How a lease rental is worked out — and checked
A lease quote is not a price list. It is a small financial model, and one wrong input can turn a lease into a loss. The CRM does the maths the same way every time and puts a person in charge of approving it.
What goes in
The car (make, model, variant from the catalogue), ex-showroom price, discount, registration, GST, tenure, interest rate and kilometres.
- Lease type: market residual value or fixed residual value, with its %
- Insurance, maintenance, accessories, hypothecation, logistics and FASTag
What the engine calculates
On-road value, the factory price without GST, and the residual value at the end of the lease.
- The monthly rental uses the standard finance formula with the residual value as the end balance
- Registration cost is spread over the tenure; the GST credit on the car is offset month by month
- Insurance and maintenance carry their own GST
- Results: monthly rental, rental without maintenance, and the net investment instalment
An approval queue with reasons
Every new quote waits for approval. Only approvers can pass it.
- A rejection needs one of ten reason codes — wrong input, wrong client terms, wrong scope, price too low or too high, missing item, wrong residual value, GST split, commercial judgement, presentation
- The reviewer can enter the right values and say “this time only” or “always for this client or model”
Rejected quotes fix themselves
After a rejection, the CRM applies the corrections, recalculates through the same engine and sends the quote back for approval — with old and new values side by side.
- If the reviewer only wrote a note, AI reads the note to work out the corrections
- It does not auto-revise price or commercial decisions, and stops if the new rental makes no sense
- If the same correction keeps coming back for a client, it suggests a standing rule
Final quote, PDF and copies
Approved quotes become a mail-ready PDF. One quote per client can be marked final.
- Marking final tells the master admin and operations straight away
- Duplicate a quote as a draft with a new client quote number
Workbooks for complex deals
For unusual deals, the team builds the quote in an in-browser spreadsheet from Tristar’s own template, with 5%, 18% and 40% GST sheets.
- One column per scenario, a minimum pre-GST income check, PDF and mail view
- Workbooks go through their own approval queue
Three to five years of every car, managed
After delivery, the real work begins. The Fleet & Ops module gives operations one place for everything that happens to a car until it leaves the fleet.
Active fleet register
Every leased car with its company, driver, dates and lease terms.
- Bulk import from Excel with a preview before anything is saved
- End-of-lease decision on each car: exit, fixed extension or holdover
Financial-year turnover projection
Actual rentals so far plus a per-car projection to March.
- Part-month rules for the delivery and exit months
- Holdover income shown separately as upside, never in the headline
Early termination
Each case tracked from proposal and calculation sheet to quote, payment, car pick-up and no-dues certificate.
- Options: buy back, surrender or not opted
Rentals, dealers, hypothecation and cess
Month-by-month rentals per company — expected, received and outstanding — with a drill-down to each driver.
- Dealer reconciliation: submitted and received dates
- Hypothecation release and RC dispatch tracking
- Cess recovery with GST and no-dues status
Daily Actions — the to-do list that writes itself
Every morning the CRM lists what needs attention.
- Leases past expiry with no decision
- Leases ending within 30 days (critical inside 7)
- Early-termination sheets pending, or paid but car not picked up
- Hypothecation release pending over 90 days; dealer reconciliation pending over 60 days
Letters in one click
RTO NOC and no-dues letters filled in from the registration number, then previewed or saved as PDF.
- Ops reports by product, state and lease action, with CSV export
Knowing what each car really costs — before pricing the next one
Maintenance and insurance decide whether a lease makes money. The CRM turns every repair bill and cost entry into pricing knowledge.
AI reads every repair bill
Upload a scan or photo and AI pulls out the registration, date, workshop, city, odometer, parts, labour, consumables, total and claim reference — with a confidence level for each field.
- Each job sorted into 13 categories: periodic service, brakes, tyres, battery, suspension and steering, body and paint, AC and cooling, electrical, engine, clutch and transmission, CNG and fuel, accident and insurance, consumables
- Low-confidence items go to a review queue; a person confirms before anything is saved
- Insurance-recoverable only when it is accident work with a claim reference
Running cost per kilometre
The calculator finds the best available cost per km for a car: the exact model first, then make and fuel, then fuel across makes, then make, segment and finally the whole fleet.
- Segments: luxury, premium, mass and commercial
- Insurance adjusted for the car’s insured value, falling each year
- Cars with impossible cost figures are left out
Pricing bands and win rates
Past approved quotes show which maintenance-and-insurance price bands win deals and which lose them.
- Needs at least 8 decided quotes before a band is shown
- A minimum margin floor, and market benchmarks that fade out as they get old
Dashboards on spend
Spend by work type, cost by make, monthly trend and the parts-labour mix, per car, per km or in total.
- A heatmap of spend by make, exportable to Excel and PDF
- Insurance and maintenance summaries per car and per make
Accounts without re-typing
Tristar keeps its books in TallyPrime, and that is not going to change — nor should it. So instead of asking accounts to export and upload files, the CRM talks to Tally directly through Tally’s XML gateway and brings in ledgers, vouchers, cost centres and balances. Finance then sees the year’s P&L and balance sheet (with a check that group totals add up), can browse and drill into every voucher and ledger, and sees insurance and maintenance per car from Tally’s cost centres — all inside the CRM. It is the kind of Tally-linked business software we build for companies that keep their books in TallyPrime.
This was the hardest engineering in the project. A few things we had to solve:
Tally only answers when the company is open
Normal report requests come back empty if nobody has the company open on the Tally PC. We switched to requests that make Tally load the data from disk, so the sync works at night too.
Tally’s XML is not always valid XML
It can contain control characters and undeclared tags that crash a normal parser, so every response is cleaned before it is read.
Big requests make Tally drop the connection
Data is fetched one financial year at a time, with retries and stall detection. If a fetch fails half-way, the old data stays untouched rather than being half-replaced.
Totals must match Tally exactly
Adding up voucher lines can silently miss postings, so every reported figure is taken from Tally’s own ledger balances. When we checked, 523 of 530 ledger balances matched the client’s Tally exports to the rupee; the other 7 were vouchers entered after the export was taken.
Ledgers must roll up the full group tree
An early version only looked one level up and missed debtors and creditors filed under sub-groups. Fixing the roll-up brought back hundreds of missing ledgers.
Tristar names one Tally cost centre per car registration number. We matched those to the CRM fleet on letters and digits only (so “MH-01 AB 1234” and “MH01AB1234” are the same car): 361 of 381 cars matched automatically, and the rest were flagged for a person to check. That is what makes per-car profit possible — CRM data and Tally costs finally sit side by side.
AI reads and drafts — a human approves
We used AI only where it removes real typing and reading, and never where it could quietly change a price. These are the three places it runs today, built on the Claude AI API. (This is the same approach we take in our AI implementation work.)
- It never makes up a number — every figure comes from the proforma, the rules or past data
- It cannot change a price taken from the dealer’s proforma
- It refuses quotes that would be loss-making or break the pricing caps
- It cannot skip approval — a person signs off every quote
AI quote assistant
Inside the quote form there is a chat assistant. A sales person drops in a dealer’s proforma invoice (PDF or photo). The assistant reads it, finds the exact make, model and variant, finds the lead, looks up past pricing and Tristar’s pricing rules, and prepares the quote. It then goes into the same approval queue as any other quote.
AI repair-invoice reading
Workshops send repair bills as scans and phone photos. The CRM reads them with AI and pulls out every line item — part, labour, amount — then sorts each line into warranty, accident, tyres and battery, or regular wear, and works out who should bear the cost. A person reviews and confirms before anything is saved. More than 350 repair invoices and about 3,000 line items have gone through this.
AI quote correction
When an approver rejects a quote with only a written note, AI reads the note, works out which values to change, and the CRM recalculates the quote and sends it back for approval. Price and commercial decisions are never auto-revised.
AI data-health check
A scheduled data-health agent checks the CRM for problems: impossible values, cost entries with an extra zero, records that don’t reconcile, deals stuck in a stage, and dashboard numbers that don’t add up. It learns the normal cost range for each car make, so an odd bill stands out. Problems become flags: critical ones are emailed straight away, warnings go out in a daily digest, and a badge in the menu shows what is open. Managers can also ask it questions in plain English; it answers using read-only database access that cannot change data or see passwords. It has raised more than 2,700 flags so far.
Running your business on spreadsheets and Tally?
We can map your workflow and show you what a CRM linked to Tally — with AI doing the typing — would look like for you.
What runs through the CRM
The numbers below are counted from the CRM itself (August 2026), not estimated:
Just as important is what changed day to day: one place to see any car’s full story, quotes that cannot skip approval, finance numbers that match Tally, and far less manual typing of invoices and costs.
Six automatic checks that keep the numbers right
A CRM is only as good as its data. On a schedule, an agent runs six kinds of checks and opens a flag for anything wrong — and closes it again when it is fixed.
1. Hard rules
Amounts above sensible limits, zero or negative values, a lease ending before it starts, a delivery before its order, missing PO fields, wrong state codes, impossible resale %.
2. Reconciliation
Totals that must agree do agree — lead sources against active leads, fuel types against POs, residual-value types against the active fleet.
3. Stale pipeline
Leads that have stayed in a stage longer than its time limit.
4. Dashboard audit
The dashboard figures are recomputed and compared, so a wrong chart is caught before anyone acts on it.
5. Learned ranges
For each car make it learns the normal cost range and flags bills far outside it.
6. Tally checks
Warns if Tally has not synced for 48 hours, and compares Tally billing with the contracted rental.
The whole business on one screen
The dashboard opens with what management asks about every week, and the reports go one level deeper.
04Dashboard — sales
- Headline numbersValue booked, POs booked, delivered, cancelled, active leads, active fleet, pending POs, companies
- Booking momentumThis month at a glance
- Sales pipelineLead-stage funnel and lead sources
- Open pipelineUn-allotted, allotted and disbursed POs
04Dashboard — fleet
- Orders by make and fuelPlus PO status breakdown
- Fleet distributionBy make and by state
- Insurance and maintenancePer make, lifetime or total, and as a % of ex-showroom
- Used-car salesAverage resale %, best and worst holding make, resale by age
05Reports
- Leasing and lease bookActive leases by company and driver
- Quote reportAll quotes with filters, plus quotes missing a price
- Account reportRentals received, with entries
- Residual value performanceUsed-car P&L against the residual value
- Insurance, maintenance and repair summariesPer car and per make
Seven roles, one shared workspace
Sales, operations, accounts, after-sales and RTO all work in the same system — each seeing only what they need.
Seven roles, fine-grained rights
Master Admin, Sales Admin, Sales, Operations, After Sales, RTO and Accounts, with a permission grid of modules and actions for every user.
- The master admin can switch into any user’s view to see what they see
- Every action recorded in an activity log, filterable by user, module and action
Task — team chat built in
Channels and direct messages, threaded replies, file sharing, @mentions and unread badges.
- Deal discussions stay next to the deal, not in personal WhatsApp
Account checklist per deal
Dealer, product, lease or outright, tenure, kilometres, fixed residual value, who handles insurance, maintenance and registration, the amount payable to the dealer, and who authorised it — saved as a PDF.
Backups the admin controls
Scheduled database backups kept for 30 days, plus a one-click download, for the master admin only.
Built to last, owned by the client
Laravel 13 on PHP 8.3 with MariaDB; Tailwind CSS and Chart.js for the dashboards; Luckysheet for in-browser spreadsheets; TallyPrime’s XML gateway for accounts; and the Claude AI API for the quote assistant, invoice reading and data-health agent. Seven user roles with per-module, per-action permissions, a full activity log, and built-in database backups — and we keep it running under our maintenance and support plan.
Four rules we would use again
Build around the workflow, not a template
Leasing has steps no generic CRM has — residual value, hypothecation, cess, end-of-lease decisions.
Keep Tally as the books
Integrate with the accounting system people trust instead of trying to replace it.
AI should read and draft, people should approve
Every AI step here ends with a human sign-off, and the AI is blocked from touching prices.
Check the numbers against the source
We tested Tally figures ledger by ledger before anyone relied on them.
If you run a leasing, rental or fleet business — or any business where one deal has a long life after the sale — see our software development and AI automation services, or read how we built a daily sales reporting system for a 100-person sales team.
What we can build for you
Frequently asked questions
What does a car leasing CRM do?
A car leasing CRM tracks each lease from the first enquiry through quote, approval, purchase order, KYC, delivery and the years of rentals, repairs and insurance after it, up to the end-of-lease decision. Tristar’s CRM also links to TallyPrime so finance figures sit next to each car.
Can a CRM be integrated with TallyPrime?
Yes. Tristar’s CRM reads ledgers, vouchers, cost centres and balances from TallyPrime through its XML gateway. Figures are taken from Tally’s own ledger balances so they match the books, and each car is matched to its Tally cost centre by registration number.
How is AI used in the Tristar Leasing CRM?
AI drafts quotes from a dealer’s proforma invoice, reads scanned repair invoices into line items, and runs a data-health check that flags mistakes. In every case a person reviews and approves, and the AI is not allowed to change prices or skip approval.
Why build a custom CRM instead of using Salesforce or Zoho?
Leasing has steps that general CRMs do not cover — lease quotes with GST and residual value, hypothecation, cess recovery, end-of-lease decisions and per-car profitability linked to Tally. A custom CRM fits those steps exactly, and the client owns the code and data.
Does Infoways build CRMs for other industries?
Yes. We build custom CRMs, dashboards and AI automation for businesses across India, including a daily sales reporting system for a 100-person sales team. Contact us for a free consultation.
Want a system like this for your business?
Tell us how your team works today. We’ll map it and show you what a custom CRM, dashboard or AI automation would look like — with an honest estimate. No cost, no obligation.
- Honest advice, no pushy sales
- Fixed, transparent pricing
- A real Mumbai team on call
