Equity research · AI that cites its sources

Institutional-grade research, with AI that shows its work.

Avanis AI extracts management guidance from earnings calls and filings with verbatim citations, keeps score on whether management delivers what it promises, and exports valuation models you can defend, number by number.

Get Avanis AI See the platform

Formerly StockLens. Impersonal research analytics, not investment advice. Prices in USD.

Avanis AI · Kaynes Technology · Q1 FY27 concall
What did management guide for FY27 revenue, and do they usually deliver?
FY27 revenue guidance ₹4,500 cr +60% YoY
“…we remain confident of closing FY27 at about ₹4,500 crores of revenue on the back of the order book…”
CFO · concall transcript · 00:41:57 · verbatim, click to open
Revenue guidance: delivered 7 of last 9 quarters
Answered instantly from stored analysis, updated after every new filing.
› ask about margins, capex, order book… Ask
Illustrative example of Avanis AI output.
What it replaces

The work you stop doing by hand.

Digesting a quarter of disclosures

Before: listen to the call, read the deck, take notes. Repeat for every prior quarter you need.

With Avanis AI: guidance, numbers, and commentary extracted from this call and every past one, compared quarter over quarter, all cited.

Checking promised vs delivered

Before: a hand-maintained spreadsheet nobody fully trusts.

With Avanis AI: management credibility measured across every number they guided, reconciled across concalls, presentations, and filings, quarter after quarter.

Building the valuation

Before: retyped numbers, lost sources, "where did this come from?"

With Avanis AI: a formula-linked model with financials pre-filled and assumptions drafted from stored guidance and credibility data. You adjust them to your thesis. It stays your model, every input cited.

01

Platform

Guidance extraction, credibility scoring, investment memory, and valuation exports. Every number traced to its verbatim source.

See the platform →
02

Tiers

Free to try, expense the Analyst tier, or bring the whole desk. Prices in USD. Join the waitlist to get in.

See pricing →
03

Research

Notes and breakdowns from the same machinery the product runs: how we read calls, score guidance, and build models.

Read the research →

Firepower for the teams that compete on ideas.

Hong Kong is full of funds competing against firms many times their size. A smaller team can’t match a big fund’s headcount, and doesn’t need to: its edge is the thinking. We build to take everything else off the desk.

Why we built it →
Team

Built from inside investment research.

Yash Dutt, Founder and CEO
Yash Dutt
Founder & CEO
Equity research and automation across boutique investment firms; AI & Automation at EY; Head of Equity Research at I-Circle. Leads product vision and finance domain.
yash@avanisai.com
Saanvi Shukla, Chief Technology Officer
Saanvi Shukla
Chief Technology Officer
Deloitte Hong Kong AI & Data Consulting; Y Combinator AI Startup School. Leads AI architecture, infrastructure, and backend systems.
saanvi@avanisai.com
Antariksh Verma, Head of Engineering
Antariksh Verma
Head of Engineering
Published research in leading computer science venues; 15+ hackathon wins. Leads platform engineering and product development.
antariksh@avanisai.com
Anish Khanna, Chief Operating Officer
Anish Khanna
Chief Operating Officer
Equity and alternative investment research at Kaleidoscope Capital. Leads operations, partnerships, and go-to-market.
anish@avanisai.com
Values

What we optimize for.

Auditability

If a number can’t be traced, we don’t show it.

Evidence over opinion

Scores are statistical measures with published methodology, never calls and never targets.

Honesty about uncertainty

When the model isn’t sure, it flags. It never silently guesses.

Questions analysts ask first.

No. Avanis AI is a research analytics tool: impersonal, not tailored to any person’s circumstances, with no recommendations, ratings, or price targets. Avanis AI is not registered with the SEC, SEBI, or SFC as an adviser, broker-dealer, or in any other capacity. You are responsible for your own investment decisions.

Avanis AI is built to save you research hours, not to promise investment returns. It pulls transcripts, filings, and analyst commentary into one place, flags where management tone diverges from the numbers, and lets you cover more names in the same amount of time. The ROI shows up as faster reviews and wider coverage. What you do with that edge is your call.

No. We never use your data or queries to train models. Position sizes and trade intent are stripped before any request leaves our systems. Product analytics (which features get used) are kept separate from your research content. Details are on our Security page.

At launch: United States (SEC filings and earnings-call transcripts, starting with US-listed ADRs), India (BSE concall transcripts and filings), and China (US-listed ADRs and Hong Kong listings first).

Analysis runs when new documents publish and is stored, so companies open instantly, already up to date. You never wait for a re-run or work from stale output.

Yes. Valuation models export as formula-linked spreadsheets with live formulas; open them, trace them, change them. Every input carries its citation.

An API is planned for the Desk tier and above. It is not available yet.

Yes. The Desk tier (3 seats minimum) is built for teams. Shared watchlists, a team workspace, a team view for the CIO, API access and SSO are planned and not yet available. Larger teams get volume pricing.

Yes, two ways: the free Explorer tier (3 companies, no card) and a 14-day money-back guarantee on paid self-serve tiers.

Earnings-call transcripts, exchange filings, and investor presentations, pulled from public statutory sources as they publish. Every extracted number cites the exact document and location it came from.

It says so. Extractions below the confidence threshold are flagged for review instead of shown as fact. Nothing is silently guessed.

Talk to us.

Coverage requests, team pricing, or an Institutional conversation. We reply within 24 hours.

NDA-first engagements On-prem on request We never train on your data
Book a meeting Pick a time that works for you. Added straight to both calendars. →
or send a message

Prefer email? yash@avanisai.com