Traceable
Every number links to the document, page and table it came from. No source, no number.
Revenue LKR 48.2bn · AR FY25 p.112
Noēsis turns core-system data, company filings and market prices into figures and models you can trace to their source, for liquidity and risk teams, research desks and the committees they answer to.
Every number links to the document, page and table it came from. No source, no number.
Revenue LKR 48.2bn · AR FY25 p.112
Nothing reaches a committee or board unverified, and its review status is always shown.
◉ Verified
Every model shows its assumptions, its version and its date, one click from the result.
Base case assumes 6% volume growth
Models estimate; they do not know. We show ranges, and what would change the view.
Model value LKR 124–176, base 151
Each product does one job well and shares the same foundations: sourced data, versioned models, a review trail and approvals named by role.
Maturity ladders, liquidity ratios, stress tests, early-warning indicators, and board and regulatory reports, built from your core-system extracts.
For: Finance companies, banks
Company research, valuation models with ranges, and a committee workflow from first draft to approved decision.
For: Brokers, investment managers
Verified financials for every company listed on the Colombo Stock Exchange, with screening, comparisons and exports.
For: Research teams, funds, insurers
Repricing gaps, and earnings and value sensitivity to interest-rate moves.
For: Finance companies, mid-sized banks
Funding plans, internal funds transfer pricing and product profitability.
For: Mid-sized banks
Recovery-plan indicator dashboards and a scenario library.
For: Banks, finance companies
Built for finance companies and banks that need to measure, monitor and report liquidity risk with an audit trail, without a global vendor's price or a fragile spreadsheet.
Illustrative figures. A model value is a model estimate, not a recommendation.
Your analysts should spend their time on judgement, not on retyping annual reports. Noēsis gives your team a clean, consistent engine to build on.
The review mark appears on every figure set, model, note and return. Nothing reaches a committee, a board or a regulator without passing through each stage.
Data arrives from your core systems, filings or exchange feeds, stamped with its source and date.
An independent reading checks each figure against its source. Differences are flagged, never overwritten.
An analyst or risk manager reviews the figures and models, and the review is signed and dated.
The responsible body (investment committee, ALCO or board) approves, and the decision is recorded with its evidence.
Noēsis uses statistical and machine-learning models where they help, and presents their output as estimates with ranges, never as predictions.
Regulated institutions must assess the technology they rely on. Noēsis is built to make that assessment straightforward.
Every change, review and approval is logged with who, what and when. Nothing is deleted or silently overwritten.
Access by role, with separate preparers, reviewers and approvers, so four-eyes control is built in.
Client data is kept separate and is never used to serve another client. House views stay in your private overlay.
Exchange and third-party data are displayed, stored and processed only as each licence permits.
We provide the documentation your outsourcing and technology-risk reviews require, including hosting, security and exit arrangements.
Nothing reaches a committee, board or regulator without a named person's approval.
Liquidity risk frameworks, stress testing and regulatory reporting.
Liquidity · ALM · RecoveryAsset-liability management, funding plans and funds transfer pricing.
ALM · Funding · RecoveryResearch teams that need verified data and models across the market.
Research · DataUnit trust and portfolio managers, insurers and pension funds.
Research · DataConsistent, English-language coverage of Sri Lankan listed companies.
Data · ResearchWe are selecting a small number of institutions for three-month pilots starting in the first quarter of 2027.