AntHill · Decision Intelligence

Org. Brain, helping transform
data into decisions
in minutes not days

AntHill is the living Context Graph that listens to every Slack, JIRA, Git, data definition, and decision, with complete understanding of your data tables.

Request access → Read the product note →

StageDesign Partnerships

MarketUS, UK, APAC · FinTech & BFSI

ShippingJuly 2026

The Problem

The most expensive workflow in your company is the diagnostic question.

Every week, your analysts run 30 to 40 diagnostic cycles. Why did activation drop? Why is reconciliation off? Why did the funnel break? Each one takes a few days — not because your analysts are slow, but because more than 60% of their time is spent reconstructing context across Slack, JIRA, Confluence, and Git before any analysis begins.

The answer existed in your organization before they started looking. They just had no way to find it.

Time to answer A few days Per diagnostic question

Cycle frequency 30–40 Cycles per week, per analytics team

Revenue exposure Millions Of dollars, compounding while answers wait

The Insight

The model isn't the moat.
The context is.

Every enterprise has frontier AI. None have organizational context.

GPT, Claude, Gemini — capable, cheap, improving every quarter. The extraction layer is solved. What none of them have is your organization's tribal knowledge: the metric definition that changed last October, the engineering incident that explains last month's drop, the prior decision buried in a Slack thread nobody will ever search.

Intelligence without context produces confident wrong answers. In enterprise analytics, a confident wrong answer is worse than no answer.

The application layer — chat, SQL generation, dashboard summaries — will commoditize within 18 months. Infrastructure won't.

AntHill is infrastructure.

The Architecture

Three layers. One brain.

AntHill sits between your enterprise data and every AI tool your team uses — permissioned, temporal, auditable, org-specific.

01 · Context Graph

The living map of how your company works.

Bi-temporal knowledge graph ingesting Slack, Confluence, JIRA, and Git. Operational in 2 weeks. Compounds with every query.

02 · Ontology Layer

The truth layer for analytics.

Metric definitions, table mappings, business rules — auto-populated from your existing dashboard SQL and Git history. Grounded text-to-SQL without hallucination.

03 · Multi-Agent Pipeline

Specialized agents that decompose, retrieve, query, validate.

Specialized agents that decompose questions, retrieve context, generate and execute queries, validate outputs. Human-in-loop at every meaningful checkpoint.

See the full architecture and use cases →

Proof

Built by operators who ran analytics at a Bank valued at $1Bn+.

AntHill exists because we lived the problem. Our founding team led analytics, product, and risk at a multi-product fintech operating at $1B+ scale — where a 50-person analytics team spent half its capacity hunting context instead of solving problems.

We built a lightweight, markdown-driven predecessor inside that company — a set of internal configs, not a product. Fifty analysts now use it daily. Six months of compounding usage proved the strategy works.

AntHill is the rebuild — same insight, a grounds-up rebuilt, enterprise-grade, shipping July 2026.

Active usage 50+ Daily analysts on the predecessor system

Time-tested 6 months Of compounding usage validating the architecture

Time recovered 30% Of team-level work compressed from days to minutes

"The context graph compounds. By month six, a competitor starting fresh at the same customer starts at zero. AntHill is already irreplaceable."

— ANTHILL PRODUCT NOTE · JULY 2026

Philosophy

Ants don't have managers.
They have context.

Alone, an ant works. Together, they move worlds.

Read the manifesto →

Work with us

We're taking on a small number of design partners through 2026.

AntHill is shipping enterprise-ready in July 2026. We're working with a small group of design partners who want their analytics function operating at thinking speed — and who are willing to shape the product as we build it.

You're a fit if

You're a fintech, NBFC, payments, or BFSI company at $10M+ revenue
You run a 20+ person analytics function across multiple product lines
Your team actively documents in Slack, Confluence, and JIRA
You have an AI-receptive senior leadership team
You feel the cost of slow diagnostic cycles every month

What you get

Direct access to the founding team
A live Context Graph operational within 2 weeks of integration
The first three months of usage shaped around your highest-value workflows
Pricing built for design partnership, not procurement

Annual contracts · Seat-based with platform fee · Custom pricing for design partners

Product · Decision Intelligence

The context layer between
your data and every AI tool
your team uses.

AntHill is three things working as one: a Context Graph that maps your organization, an Ontology Layer that grounds every query, and a Multi-Agent Pipeline that does the work — with human review at every meaningful step.

The Architecture

Three layers. One brain.

01

The living map of how your company works.

The Context Graph ingests Slack, Confluence, JIRA, and Git — your organization's actual record of how decisions get made, what changed, and why.

It's bi-temporal. Permission-aware. Multi-tenant from day one. Sixteen entity types. Tiered extraction. Full entity resolution. Operational in 2 weeks of integration.

Every query enriches the graph. Every analyst correction strengthens it. By month six, a competitor starting fresh at your organization starts at zero.

Bi-temporal · 16 entity types · Permission-aware · Multi-tenant · 2-week activation

02

The truth layer for analytics.

Every analytics team has tribal knowledge about its data: which table is canonical, which filter excludes which edge case, what the metric definition was before the change last October.

The Ontology Layer captures all of it. Auto-populated from your existing dashboard SQL, Git history, and analyst-validated query library. Continuously tuned during the first weeks of usage.

This is what makes text-to-SQL work without hallucination — because the ontology grounds every query in your team's agreed-upon definitions.

Metric definitions · Table mappings · Business rules · Trusted query library

03

Specialized agents that do the work.

A diagnostic question gets decomposed into structured hypotheses. Specialized agents work in parallel — context retrieval, hypothesis ranking, query generation, validation, inference, decision-trace documentation.

Every step is visible to the analyst in real time. Every output is cited back to its source. Human-in-loop at every meaningful checkpoint — not because we're cautious, but because judgment is the product.

The agents do the execution. The analyst does the thinking.

Hypothesis generation · Grounded query execution · Cited outputs · Human-in-loop

What changes

From a few days to just a few minutes.
For 80% of diagnostic questions.

Same analysts. Same data. The work that used to fill a week now happens between meetings.

Before AntHill

A Slack message arrives Monday at 9am. Why did activation drop?

An hour confirming the drop is real. Twenty hypotheses arriving from PMs, ops, leadership. Two to three hours per hypothesis reconstructing context across Slack, JIRA, Confluence, Git. SQL written from scratch. Output interpreted manually. One hypothesis validated. Repeat.

A few days to an answer. Then more weeks to act on it.

With AntHill

The analyst types the question. The Context Graph has already indexed every thread, ticket, document, and query. Specialized agents decompose the problem, rank hypotheses by plausibility, run validation queries against the Ontology, return cited answers in parallel. The analyst reviews, corrects, confirms.

Root cause identified, cited, and documented in minutes. The decision happens before lunch.

The analyst stops doing archaeology.
They start doing analysis again.

Use Cases

Four real workflows. Four personas. One platform.

Analyst · Metric moves in the wrong direction

The Diagnostic Investigation

A few days of context-hunting → one morning of confirmed answers. The Context Graph indexes every thread, ticket, and query. Agents decompose, rank, validate, document. The analyst reviews and confirms.

Outcome: Five days to one day. For 80% of diagnostic questions.

Analytics / Finance · Routine reconciliation

The Operational Integrity Check

Payment reconciliation across issuer, payment provider, receiver, merchant. Standard cycle: 4 to 5 hours. Daily. AntHill retrieves the schema, identifies tables, generates and executes the logic, surfaces discrepancies with citations.

Outcome: 4 to 5 hours to 15 minutes. Daily.

Analyst · Optimization request with 8–10 constraints

The Complex Problem Build

Shift sizes, total shifts, break structures, gender proportion, language splits, productivity targets. Manual logic, multiple iterations, 4 to 6 hours. AntHill interprets the problem, generates the logic, runs iterations autonomously, returns a validated solution.

Outcome: 4 to 6 hours to 30 minutes. Single session.

Head of Product · Anomaly spotted in a dashboard

The Self-Serve Business Question

Today: a Slack message to an analyst, 24 to 48 hours to an answer. With AntHill: the question goes directly to the Context Graph. Cited answer in minutes — historical patterns, prior anomalies, definition history.

Outcome: 24 to 48 hours to minutes. Self-serve.

The user never sees a black box. AntHill shows its work — tasks decomposed, sources cited, queries visible, outputs explained.

Want this for your team?

We're taking on a small number of design partners through 2026.

Become a design partner →

Manifesto

Every organization
already knows what it
needs to know.

The answer to your most expensive question exists somewhere in your systems — in a decision made last quarter, in a hypothesis tested and forgotten, in a Slack thread nobody searched. We built AntHill to close that gap permanently.

What we believe

Every organization already knows what it needs to know.
The answer to your most expensive question exists somewhere in your systems — in a decision made last quarter, in a hypothesis tested and forgotten, in a Slack thread nobody searched.
The problem was never data. You have enough data.
The problem is the gap. Between what your organization knows and what your leaders can access in the moment it matters.
That gap costs you millions. Monthly. Quietly. Compounding.
We built AntHill to close that gap permanently.
Not with more dashboards. Not with more reports. With a compass — grounded in every system, every decision, every process your organization has ever touched.
Ants don't have managers. They have context. And with that context they move mountains.
So do your people. When they have AntHill.

Why the name

A compass. Not an answer.

A compass doesn't give you answers. It orients you so you can find them.

It's trusted under pressure. Reached for when the stakes are high and the cost of being wrong is immediate. You don't reach for a compass when things are easy — you reach for it when you're lost.

That's exactly AntHill's use case.

An ant alone is just surviving but together they move mountains. Not because any single ant is intelligent — but because the colony is connected. Every grain of knowledge available to every ant who needs it.

Your organization works the same way. The intelligence already exists. AntHill connects it.

What we believe

Three convictions that shape every product decision.

01

Activation over accumulation

The data was never the problem.

The most valuable intelligence in any organization is already inside it — unstructured, disconnected, dormant. The job is activation, not accumulation. More data was never the answer.

02

Analysts deserve better

Elevate, don't replace.

Decision makers shouldn't need analysts to access truth. Analysts should be elevated above execution — into strategy, judgment, and work only humans can do.

03

Speed without grounding is dangerous

Fast and right are not the same.

The goal is never fast answers. The goal is fast, accurate, contextually grounded answers. Those are not the same thing. We will never optimize one at the cost of the other.

What we won't do

Three lines we won't cross.

×

Sell speed without accuracy

A fast wrong answer is worse than a slow right one. We will never optimize for response time at the cost of grounding. Every output is validated before it reaches a decision maker.

×

Replace human judgment

AntHill gives decision makers direction — not decisions. The compass points north. The leader still chooses to walk. We enhance judgment; we never substitute it.

×

Add complexity in the name of features

Every addition to AntHill must make the decision maker's experience simpler, faster, and more certain. Complexity is the enemy of the compass. If it makes the product harder, it doesn't ship.

How we build

Six principles that guide every release.

P.01

Grounded or silent.

Every answer traces to a real source. If the context doesn't exist, AntHill says so. We don't fill the gap with inference.

Excludes: Output not grounded in indexed organizational context.

P.02

Human judgment is the product.

The human-in-loop is not a safety feature bolted on. It's the architecture. At every meaningful checkpoint, the analyst reviews, corrects, confirms.

Excludes: Fully autonomous agent execution on consequential decisions.

P.03

Reliability before features.

When forced to choose between a new workflow and making an existing one more reliable, we choose reliability. Every time.

Excludes: Feature expansion at the cost of core workflow reliability.

P.04

Show the work. Always.

Every task decomposed, every source retrieved, every query written, every inference drawn — visible in real time. Transparency is how trust is built.

Excludes: UX patterns that obscure how an answer was reached.

P.05

Complexity is the enemy.

Every addition must make the analyst's experience simpler. A compass that requires a manual is not a compass.

Excludes: Features that add capability at the cost of clarity.

P.06

AntHill adapts to the organization. Never the reverse.

Analysts stay in the tools they already use. The Context Graph ingests from systems the org already runs. AntHill learns the language — not the other way around.

Excludes: Onboarding requirements that ask analysts to adopt new habits before they experience value.

Decisions at thinking speed.

Become a design partner →

Pricing · 2026

No tiers.
No public price.
Not yet.

We're shipping with 5 design partners through 2026. Pricing is shaped to that relationship — not a SKU sheet, not a procurement form, not a calculator that asks how many seats you have before it tells you anything useful.

What a design partner gets

One commercial shape.
Three things in it.

An annual platform commitment. Single contract, single AntHill instance, multi-tenant inside your org. No per-feature add-ons, no usage-tier thresholds that trigger renegotiation mid-year.

Seats for the analysts using it. Priced per active analyst seat, scoped to the teams in the rollout. We size the seat count with you against the workflows in scope — not the headcount of every adjacent team.

The first three months as joint work. Implementation, ontology calibration, and the first three production workflows are co-built with our team. Not a services line item — part of the deal.

Public pricing arrives when general availability does. Until then, the conversation is the quote.

Want a number for your board?

Tell us the team and the workflows. We'll come back within a week.

founders@theanthill.ai →

Docs · Private preview

The docs live
inside the product.

AntHill is in private preview with design partners. Reference docs, architecture deep-dives, and the integration runbook are gated to deployed orgs — they evolve weekly with the product, and a public copy would be wrong by the time you read it.

What we will share today

Three things,
on request.

The product note (PDF). Pages on the architecture, the workflows, and the trade-offs. The same document the founding investors read. Download (available for partners) →

An architecture walkthrough. A 45-minute call with the CTO/CPO/CDO covering the Context Graph, the Ontology Layer, and the agent pipeline at the depth your eng team needs. Booked one-to-one.

The security packet. SOC2 readiness status, threat model, multi-tenancy boundaries, KMS posture, and the responses to the procurement questionnaires we've already answered. Sent under NDA.

Public reference docs ship alongside general availability in 2026. Sign up below and we'll mail you the index the day they go live.

Want one of the three?

One email. Tell us which. We'll route it.

founders@theanthill.ai → Read the product note →

About

We built the system
we wished we had.

AntHill's founding team comes from analytics, product, risk, and engineering at a multi-product fintech valued at $1Bn+. We spent years watching brilliant analysts spend >60% of their day hunting context across five tools — and the remaining 40% being too tired to think clearly.

The Origin

We tried to fix this from the inside.
Then we left to build it for everyone.

Better infrastructure helped — PySpark, Airflow, semantic models, dashboards. The team got 10% faster. The core problem didn't move.

In 2025, with frontier AI finally capable enough, we built a markdown-driven predecessor system inside our company. An orchestration layer with localized context and AI agents. Within four months, fifty analysts were solving 25% of their execution-heavy work in minutes instead of days. Average working day dropped from 15 hours to 10. Same output. More motivation.

The proof point was undeniable. The opportunity was bigger than one company.

AntHill is the rebuild — same insight, enterprise-grade, shipping July 2026.

The Team

Four founders. Years of working together. One conviction.

Nikhil Gundawar

CEO & Co-founder

11 years in Analytics, AI, and Data Science. Previously at slice, Fractal.ai, Asper.ai, Citibank, Mu Sigma.

Mihir Dash

CTO & Co-founder

4 years in Analytics and AI engineering. Previously at slice and FrontRow.

Aishwary Rathi (MBA, XIMB)

COO & Co-founder

11 years in Analytics and Strategy. Previously at slice, Tiger Analytics, Accenture. MBA, XIMB.

Bhavuk Manocha (IIT Dhanbad)

CSO & Co-founder

7 years in Analytics and AI engineering. Previously at slice, Flipkart and Tredence. IIT Dhanbad.

How we work

Three years working together. Three years of evidence.

The founding team has shipped together at scale. We've built core banking analytics that reduced ledger differences from 22% to under 0.01%. We've scaled lending AUM by 150% in two years. We've driven CX resolution rates from 60% to over 85%. We've deployed AI tools that recovered 60% of legal and compliance time and 30% of analytics team capacity.

We know how to build with each other. We know how to disagree productively. We know how to ship under deadline.

The product comes from the same place all of those did — operators solving problems we lived inside.

Become a design partner → See open roles →

Careers · Building through 2026

Build the brain
every company
will need.

We're hiring four founding-team engineers to ship AntHill's enterprise build by July 2026. This is the smallest the team will ever be. The work is infrastructure that compounds. The trade-offs are real. The upside is meaningful.

Why here, why now

The smallest the team will ever be.

01 · Infrastructure that compounds

Foundations, not features.

You'll build the Context Graph, the Ontology Layer, the agent pipeline — the systems every model in the next decade will need to run inside an enterprise. Not features. Foundations.

02 · Direct work with founders

Code reviewed by the CTO.

Four co-founders. Eleven years working together at $1B+ fintech scale. You'll ship code reviewed by the CTO, see your work go to design partners within weeks, and own systems end-to-end without committee.

03 · Real equity, real trade-offs

The honest deal.

Pre-seed equity, sub-market salaries, high pace. The honest deal: meaningful ownership, real ambiguity, and a team that has shipped together before. Not for everyone.

Open roles

Four roles. All founding team.

Backend Engineer

Bangalore · Hiring 2

The work is the architecture. The Context Graph and the Ontology Layer are the moat — you'll build them.

What you'll build

Bi-temporal knowledge graph in PostgreSQL with pgvector — recursive CTEs, HNSW indexes, RLS for multi-tenancy
Async ingestion pipelines from Slack, Confluence, JIRA, Git via Celery + Redis with tiered freshness
MCP-native tool layer over FastAPI — auth, permissions, audit logging at the tool boundary
Entity resolution and tiered extraction at enterprise scale (200K+ nodes per org)

Python · FastAPI · PostgreSQL · pgvector · Celery · Redis · MCP

careers@theanthill.ai →

GTM Engineer

Bangalore or Remote · Hiring 1

Build the systems that turn warm intros into design partners. GTM at AntHill is software, not slides.

What you'll build

Outbound infrastructure — sequencing, attribution, response handling — owned end-to-end
CRM and pipeline automation tied to product usage signals from the Context Graph
Onboarding telemetry: which design partners are activating, which are stuck, which signals predict expansion
Sales engineering tooling — demo environments, integration runbooks, security review materials

Python · TypeScript · CRM APIs · Segment · SQL · Automation tooling

careers@theanthill.ai →

Security Engineer

Bangalore · Hiring 1

Enterprise contracts hinge on this work. You'll own SOC2 from zero and build security into the architecture, not on top of it.

What you'll build

SOC2 Type I readiness — controls, evidence, audit infrastructure — from the ground up
Permission system architecture: row-level security, ACL synchronization, fine-grained audit logging
Threat modeling for the Context Graph — sensitive data ingestion, multi-tenancy, key management
Security review responses for enterprise procurement — runbooks, attestations, technical documentation

PostgreSQL RLS · Identity Providers · SOC2 · KMS · Observability · SAST

careers@theanthill.ai →

How we work

Honest about the trade-off.

Salaries are below market. Equity is meaningful. Hours are real — not 996, but startup pace, and the first six months will demand more than the next eighteen. We work primarily from Bangalore with regular travel to the US. Founders are accessible. Ambiguity is constant. Every founding-team hire shapes the company they joined.

If you want maximum optionality and minimum risk, this is not the place. If you want to build infrastructure that compounds — on a team that has shipped together before — apply.

Apply

One email. No forms.

Send an email to careers@theanthill.ai with the role title in the subject. Include:

· A short paragraph on why this role specifically

· Your most recent work — code, writing, project, anything that shows how you think

· A LinkedIn or CV

We respond to every application within seven days. If we want to talk, the first call is with a founder.

careers@theanthill.ai →

Org. Brain, helping transform data into decisions in minutes not days

The most expensive workflow in your company is the diagnostic question.

The model isn't the moat.The context is.

Three layers. One brain.

The living map of how your company works.

The truth layer for analytics.

Specialized agents that decompose, retrieve, query, validate.

Built by operators who ran analytics at a Bank valued at $1Bn+.

Ants don't have managers.They have context.

We're taking on a small number of design partners through 2026.

The context layer betweenyour data and every AI toolyour team uses.

Three layers. One brain.

The living map of how your company works.

The truth layer for analytics.

Specialized agents that do the work.

From a few days to just a few minutes.For 80% of diagnostic questions.

Four real workflows. Four personas. One platform.

The Diagnostic Investigation

The Operational Integrity Check

The Complex Problem Build

The Self-Serve Business Question

Want this for your team?

Every organizationalready knows what itneeds to know.

A compass. Not an answer.

Three convictions that shape every product decision.

The data was never the problem.

Elevate, don't replace.

Fast and right are not the same.

Three lines we won't cross.

Sell speed without accuracy

Replace human judgment

Add complexity in the name of features

Six principles that guide every release.

Grounded or silent.

Human judgment is the product.

Reliability before features.

Show the work. Always.

Complexity is the enemy.

AntHill adapts to the organization. Never the reverse.

Decisions at thinking speed.

No tiers.No public price.Not yet.

One commercial shape.Three things in it.

Want a number for your board?

The docs liveinside the product.

Three things,on request.

Want one of the three?

We built the systemwe wished we had.

We tried to fix this from the inside.Then we left to build it for everyone.

Four founders. Years of working together. One conviction.

Nikhil Gundawar

Mihir Dash

Aishwary Rathi (MBA, XIMB)

Bhavuk Manocha (IIT Dhanbad)

Three years working together. Three years of evidence.

Build the brainevery companywill need.

The smallest the team will ever be.

Foundations, not features.

Code reviewed by the CTO.

The honest deal.

Four roles. All founding team.

Backend Engineer

GTM Engineer

Security Engineer

Honest about the trade-off.

One email. No forms.

Org. Brain, helping transform
data into decisions
in minutes not days

The model isn't the moat.
The context is.

Ants don't have managers.
They have context.

The context layer between
your data and every AI tool
your team uses.

From a few days to just a few minutes.
For 80% of diagnostic questions.

Every organization
already knows what it
needs to know.

No tiers.
No public price.
Not yet.

One commercial shape.
Three things in it.

The docs live
inside the product.

Three things,
on request.

We built the system
we wished we had.

We tried to fix this from the inside.
Then we left to build it for everyone.

Build the brain
every company
will need.