HOW IT WORKS

Connect your systems. Ask in plain English.
Get answers in seconds.

Kolossus reads the systems your business already runs - Tally, CRM, inventory, Excel, even WhatsApp - and answers cross-system questions in plain English.

No data warehouse.No pipelines.No migration.
TTALLY PRIMEODBCCCRM SYSTEMREST/DBIINVENTORYSQLEEXCEL SHEETSFILEKKOLOSSUSCRM × TALLY · 60+ DAYS9.2s₹47.2L outstandingacross 5 customers · 3 criticalAcme Industries₹18.4L · 78dVista Manufacturers₹11.2L · 62dTTALLYCCRMIINVENTORYEEXCELKKOLOSSUSCRM × TALLY · 60+ DAYS9.2s₹47.2L outstandingacross 5 customers · 3 criticalAcme Industries₹18.4L · 78dVista Manufacturers₹11.2L · 62dApex Retail₹9.6L · 65d+ 2 more customers

Four steps from connection to answer. Each one happens once or runs continuously - you decide.

Most BI tools require months of setup before they answer their first question. Kolossus runs the four steps below in sequence and starts working within the first onboarding call.

Step 01 · Connect

Pick the systems Kolossus
should read from.

On day one, you tell us which systems you want Kolossus to read. Tally Prime, your custom CRM, your inventory module, the Excel sheets your finance team maintains. We pick the right connection method for each one - direct DB, API, ODBC, or file ingestion.

  • 01
    No new infrastructure - we connect to what you already run
  • 02
    Read-only by default - write-back is a separate permission
  • 03
    Multi-system from day one - single Tally is the easy case
Select your systemsSystem Picker
T
Tally Prime
ODBC
C
Custom CRM
REST
I
Inventory
SQL
E
Excel sheets
FILE
W
WhatsApp
Idle
P
Postgres
Idle
4 SYSTEMS CONNECTEDACTIVE
Step 02 · Read

Kolossus learns your data
without anyone teaching it.

We discover your tables, columns, and relationships automatically. No documentation pack from your team. No data dictionary. Schema discovery is continuous - new fields, renamed tables, evolving vendor models all re-map on their own.

  • 01
    Auto-discovery across PostgreSQL, MySQL, MongoDB, SQL Server, Oracle
  • 02
    Tally vouchers, ledgers, stock entries read in their native structure
  • 03
    Field-level access controls - hide sensitive columns from queries
Schema mappedAuto-discovered
CustomersTally
customer_idint
namevarchar
balancedecimal
last_invoicedate
LeadsCRM
lead_iduuid
companyvarchar
tally_customer_idint · FK
last_contacttimestamp
StockInventory
skuvarchar
warehousevarchar
quantityint
Step 03 · Query

Ask in plain English.
Get a cross-system answer.

Type your question the way you'd ask a colleague. Kolossus decomposes it into structured queries against each connected system, runs them in parallel, joins the results, and delivers the answer with the underlying data trail visible.

  • 01
    English, Hindi, Gujarati, Tamil, Marathi - all handled natively
  • 02
    Cross-system joins happen automatically - no manual data prep
  • 03
    Average response: 9.2 seconds across millions of records
Question to queryDecomposition
ASKWhich top customers are 60+ days overdue?
Decomposed by Kolossus
CRMGet top 50 customers by lifetime value
TallyGet outstanding invoices > 60 days
JoinMatch by customer_id, rank by amount
Step 04 · Act

The answer isn't the end.
It's the start of the action.

Most BI tools stop at the chart. Kolossus shows you the answer, then offers to do something about it - write back to Tally, send WhatsApp, draft emails, update CRM. Every action requires explicit approval. Every action is logged.

  • 01
    Write-back to Tally, CRM, custom systems with permission control
  • 02
    Automated weekly reports sent on schedule, in your format
  • 03
    Full audit trail of every read and write, exportable
Answer + actionsApproval gated
CRM × TALLY · 60+ DAYS9.2s
₹47.2L outstanding
5 customers · 3 critical (90+ days)
Available next actions
Update lead statusCRM
Draft reminder emailGmail
Pin to dashboardKolossus
Notify collections teamWhatsApp
Deployment

From your servers or ours. Same answers either way.

Most AI analytics platforms run only in their cloud. Kolossus runs in our cloud or fully on-premise - your choice, capability identical.

Where the AI lives

Two deployment paths. Same Kolossus.

Indian mid-market businesses have a real cloud-versus-on-prem decision - driven by data residency, regulatory requirements, internet reliability at plant locations, and IT preferences. We support both, with feature parity.

Regulated

On-premise via Nano LLM

Air-gapped · Your hardware
Air-gappedYour hardware10.4s avgOffline-capable

The full Kolossus stack - including our Nano LLM - runs entirely inside your infrastructure. Zero data egress, your IT team has root, DPDP / RBI / sectoral regulator friendly, same query interface as cloud.

Best for
Regulated industries (BFSI, healthcare, defense suppliers), plants without reliable internet, or businesses with strict data residency policies.
SAME QUERYCustomers overdue 60+ days, ranked by amount
Cloud deployment
₹47.2L outstanding
5 customers · 3 critical
AWS · MUMBAI9.2s
On-prem (Nano LLM)
₹47.2L outstanding
5 customers · 3 critical
YOUR SERVER10.4s
Identical answer · Identical capability
What we don't do

Naming the boundaries up front.

Every product page on the internet promises everything. Here's the opposite. Naming what we don't dotells you what we're confident we do well - and saves you from a surprise during your POC.

Out of scope
We don't replace your ERP.
Tally, SAP, Oracle, your custom ERP - Kolossus reads them. It doesn't substitute for them. Your team keeps using the systems they know.
WhyERPs encode years of process. Replacing them is a multi-year migration. We sit on top instead.
Out of scope
We don't migrate or copy your data anywhere.
Your data stays in its home systems. Kolossus reads, queries, writes back through your existing APIs. No data warehouse, no ETL, no second copy.
WhyEvery data copy is a leak surface and a sync failure waiting to happen. We avoid both.
Adjacent only
We don't generate filing PDFs or submit to government portals.
For RERA, GST returns, TDS filings - we prepare the underlying data so your CA's filing prep takes hours instead of weeks. The actual filing stays with your accountant or your existing return-filing software.
WhyCompliance liability and signing authority sits with your CA. We feed the inputs, not the submissions.
Out of scope
We don't write your application code.
If you want a new app or a new business workflow built, that's still your development team's job. We make their existing systems queryable in plain English.
WhyCustom code requires understanding your business logic. We answer questions, we don't ship apps.
Adjacent only
We don't build dashboards as a side project.
Power BI, Tableau, Metabase already do dashboards well. Kolossus answers questions - sometimes those answers happen to be charts, but the visualization layer isn't our job.
WhyMost teams already own a BI tool. We complement it instead of competing with it.
Product FAQ

Questions evaluators actually ask.

No. A chatbot wraps a language model around generic knowledge. Kolossus reads your specific systems - your Tally instance, your CRM, your inventory module - and answers questions grounded in your real data.

The interface looks chat-like because that's the most natural way to ask business questions. But behind the input box is a system that decomposes your question into structured queries against your actual data.

For most teams, no - and that's deliberate. BI tools are good at recurring dashboards: same chart updated daily, same KPI on the wall.

Kolossus is good at ad-hoc questions- the questions your team thinks of in a meeting that don't have an existing dashboard. "Which customers are bleeding margin this quarter?" is a Kolossus question. "Show me weekly revenue trend" is a BI tool question.

Most of our customers run both. Kolossus complements your BI tool rather than replacing it.

Snowflake and Databricks are data warehouses- they require you to move your data into them first via ETL pipelines, then run analytics on top. For large enterprises with dedicated data engineering teams, that's a sensible architecture.

For Indian mid-market - where you don't have a 10-person data team and the cost of a data warehouse implementation often exceeds the value - Kolossus is a different bet. Read source systems directly. Skip the warehouse. Get to answers in three weeks instead of eighteen months.

Yes - via on-premise deployment with our Nano LLM. The full Kolossus stack runs inside your network with no external traffic. Updates happen through scheduled offline packages.

This deployment was built specifically for plant locations with unreliable internet, regulated industries with no-egress policies, and businesses with strict data sovereignty requirements.

English, Hindi, Gujarati, Tamil, Marathi - all handled natively. Type your question in any of these languages and Kolossus translates internally to a structured query.

More importantly: Kolossus reads data in these languages too. Vendor names in Hindi, item descriptions in Gujarati, customer notes in Marathi - no English-translation layer required.

Your data is never used to train models- yours or anyone else's. Kolossus's language understanding is pre-trained; what changes per customer is only the schema mapping and access control layer.

For on-premise deployments via Nano LLM, this is straightforward: nothing leaves your environment to train anything. For cloud deployments, we contractually commit to the same: your data is queried, not absorbed.

Annual subscription. Tier depends on number of users, query volume, and whether you need on-premise deployment via Nano LLM.

Most standard mid-market customers fit our default tier. POC is free for 14 days on your actual data - no credit card required.

We don't publish exact tier prices because each deployment is sized to the customer. The honest answer is "talk to us and we'll quote a number you can plan around."

The use case pages we publish (Tally, custom CRM, manufacturing, real estate, trading) cover the majority of our customer base. If yours isn't there, it's probably still a fit.

Kolossus is fundamentally a system that reads source databases and APIs and answers questions. That's industry-agnostic. The published pages are just where we have the most pattern recognition. Tell us what you run and what you'd ask- we'll tell you honestly whether a POC makes sense.

06 · Next step

Ready to see this on your data? Two paths from here.

Recommended
Talk to us on WhatsApp
Message a co-founder directly. Tell us what systems you run and what questions you wish you had faster answers to. We'll tell you honestly whether a 14-day POC makes sense for your business.
Start WhatsApp conversation
Founder replies usually within 5 minutes
Explore by industry
Browse use cases
Industry-specific pages with operational scenarios, sample queries, and evaluation FAQs.
For TallyFor CRMsFor ManufacturingFor Trading
Browse use cases

Frequently Asked Questions

How much does Kolossus cost?

Our Growth tier (which covers most of what a 150-500 person business needs) is ₹20,000–30,000/month. That’s ₹2.4–3.6L/year.

Is that expensive? Compared to what? If your finance team spends 8 hours a week on manual Tally reports, that’s ₹2L/year in analyst time alone. And that’s before counting the decisions made on stale data.