Skip to content
Every figure traces to a comparable, or the report does not render
LegosphereReal Estate

Valuation · Filings · Deeds

A property system of record for
registered valuers

Legosphere Real Estate values a property from an adjustment grid rather than an average, reconciles three approaches into one conclusion with a written rationale for every weight, and refuses to render a sentence about title that no registered document supports. Every number in the report resolves to the comparables underneath it.

Talk to usHow a valuation is reached
  • The model writes no figures
  • No title assertion without a document
  • Adjusted comparables, never an average
  • A draft until a registered valuer signs

Three surfaces

An engine, a workspace, and the statutory work beside them

One record underneath all three: it holds the property and its evidence, the engine computes against it without being able to write prose, and the console is where a person applies judgement and signs.

The engine

The valuation engine

Deterministic arithmetic, with the drafting kept outside it. Comparables are adjusted factor by factor, three approaches produce three indicated values, and a reconciliation weights them into a conclusion. Nothing in this path is a language model, which is what makes the output defensible.

  • Every comparable adjusted for date, location, tenure, size, age, floor, frontage, view, condition and distress
  • A written rationale per adjustment, which the schema refuses to store without
  • Sales comparison, income and cost as first-class methods, each with an indicated value
  • Weights that must sum to one, each carrying its own reason in prose
  • Exact decimal arithmetic throughout — a portfolio ties to its lines to the rupee

Registered valuers · reviewers · anyone who has had a figure challenged

The valuation engine
The workspace

The valuer console

Where the judgement is applied and the signature happens. Mandates, properties, documents, comparables on a map, the grid a reviewer reads across, the three approaches side by side, review notes against a section, and a sign-off gate that will not open early.

  • An editable adjustment grid, with the sample adequacy check visible while you work
  • Approaches shown side by side against a value range, not collapsed into one number
  • Click any figure in a report and see the line, the adjusted comparables and the source sheet
  • Review notes raised against a section or a comparable, assigned, answered and closed
  • Rejected import rows listed with the row number and the reason, never dropped quietly

Valuers · analysts · partners · reviewers at a lender

The valuer console
The statutory half

Filings and statutory work

The work that sits beside a valuation and has its own deadlines. Registration and quarterly obligations with real due dates for the state in question, approvals tracked to their expiry, and duty computed against both the consideration and the notified rate with the higher applied.

  • Registration and periodic obligations generated per state, with due dates rather than reminders
  • Approvals with validity and expiry, flagged before they lapse rather than after
  • Duty computed on the higher of consideration and the notified rate, shown both ways
  • Deeds, leases, development agreements and powers of attorney against a clause schedule
  • Rent rolls and portfolio roll-ups that total identically per property and in aggregate

Compliance practitioners · developers · transaction teams

Filings and statutory work

The specification

What it is, and what it is not

The questions that actually decide this purchase are specification questions. They are answered here rather than on a second call.

What this is

A system of record for property work: mandates, properties, documents, comparables, valuations and deliverables, with a deterministic valuation engine over the top and a drafting layer that is only allowed to write prose. It is built to a regulated basis — registered valuation under the insolvency and companies frameworks, and bank panel valuation — which is what fixes the rules the rest of this site describes.

What this is not

Not a price index, not an automated valuation model, and not a title certificate. It does not form a view on where a market is going, it does not conclude a value without comparables a reviewer can read, and it does not assert that a title is clear — it shows the registered instruments and lets a valuer say so, under their own registration.

DomainValuation, statutory filing and transaction documents — for the property itself, not for the transaction ledger around it
Basis of valueMarket, fair, liquidation, distress or insurable — declared on the mandate, never defaulted, and it decides which methods are mandatory
MethodAn adjustment grid per comparable, then sales comparison, income and cost reconciled by weights that sum to one
EvidenceOwnership, tenure, encumbrance, approvals, area and age each resolve to a registered document, a title chain entry or an approval row
ProvenanceEvery rendered figure comes from a valuation line; every line carries the comparables and the source it came from
TenancyThe firm is the tenant, enforced where the query is built. A read across firms answers as though the record does not exist
MoneyExact decimals end to end, rounded once at the boundary. No floating-point number touches a rupee figure
AreaOne conversion table both the engine and the console read. A unit whose factor varies by state refuses to convert without an explicit instruction
Sign-offA partner or a registered valuer whose registration covers the asset class, with no review note left open
IntegrationThe console is one client of the API. Completion callbacks are signed, so a lender can take a valuation into their own system

Why it is built this way

A wrong valuation is not discovered. It is relied upon.

A valuation is read by people who will act on it and who have no way to check it. A lender sizes a facility against the figure. An insolvency professional distributes on it. A court takes it as the opinion of a registered professional, because that is precisely what it is. None of those readers can tell, from the document, whether the number came from eight comparables carefully adjusted or from an average of whatever was to hand.

That is what makes the expensive failures quiet ones. An unadjusted mean of four sales looks exactly like a considered rate. A sentence saying the title is clear and marketable reads the same whether a chain of registered instruments sits behind it or nothing does. A figure a drafting model wrote into a paragraph is indistinguishable, on the page, from one the arithmetic produced — until somebody asks which comparable it came from, and the answer is that it did not come from one.

So the architecture is aimed at those specifically. Figures may only reach a rendered document from a valuation line, and a number appearing in prose without one blocks the render rather than shipping. Every factual assertion about the property has to resolve to a document a reviewer can open, and where it does not, the job stops and names what is missing instead of softening the sentence — because the hedge is the part a reader skips. The comparables are adjusted individually with a stated reason per factor, and if they still disagree too widely afterwards the software refuses to conclude rather than splitting the difference.

None of that makes the valuation right. It makes it reviewable, which is the only property a valuation can actually be given by software. The judgement stays with the person whose registration is on the last page.

How it works

What happens to a mandate

Five stages between an instruction arriving and a signed report going out, each with a rule attached. The rules are the interesting part — most exist because of a way work like this fails quietly.

  1. 01

    Take the instruction

    A mandate records the client, the purpose and the basis of value, and none of the three has a default. The purpose then decides which methods are mandatory, so a lending valuation of a tenanted property cannot later be concluded on comparable sales alone.

  2. 02

    Ingest, and reject loudly

    Comparable sheets, lease schedules, construction stages and land records are parsed row by row. An unreadable row is an error naming the row and the field, never an empty result — parsed plus rejected plus duplicate has to equal what you gave it, and the rejected list is on screen.

  3. 03

    Adjust every comparable

    Each comparable is adjusted in a fixed order for date, location, tenure, size, age, floor, frontage, view, condition and distress, each with a written reason. The sample is refused if it is too small, too old, too far away, or still disagrees after adjustment.

  4. 04

    Reconcile, never average

    Sales comparison, income and cost each produce an indicated value. Weights must sum to one and each carries a rationale, and approaches that diverge beyond the threshold are refused rather than split down the middle.

  5. 05

    Gate, render, sign

    Before anything is drafted, every assertion the report will make is checked against the documents on file; unsupported means the job stops. What renders carries only figures from valuation lines. It is watermarked a draft until a registered valuer signs it.

The engine, stage by stage

What it has to satisfy

Twelve rules that constrain the software

Each one is a constraint rather than a capability, and each is written so a customer can hold the software to it. Most of them describe something it refuses to do.

  • The model never produces a number

    A figure reaches a rendered document from a valuation line or it does not appear. A number a drafting model wrote into a sentence is a blocking error rather than a paragraph to proofread, which is the only version of this rule that survives a busy week.

  • No assertion about title without a document

    Ownership, tenure, encumbrance and approvals resolve to a registered instrument, a title chain entry or an approval on file, or the report does not render. There is no override for this — not a flag, not a setting, not a senior user.

  • An unadjusted mean is not a valuation

    A raw average of comparable rates is a sanity statistic and is labelled as one everywhere it appears. The conclusion comes from a grid where every comparable carries its adjustments and the reason for each of them.

Read all twelve

Questions

What people ask first

Mostly one question in several forms: does the software decide anything? It does not, and these are the mechanisms that keep it that way.

No, and it is built so that it cannot quietly start to. The arithmetic is deterministic and the software will tell you what the adjusted comparables indicate, what each approach indicates, and what the reconciliation of those approaches comes to under the weights someone chose. Choosing the adjustments, choosing the weights, and taking responsibility for the conclusion are the valuer’s, and the deliverable is a watermarked draft until a registered valuer signs it.

Point it at a valuation you have already defended

The useful conversation starts from a report that came back with questions rather than from a deck: which figure was challenged, what the reviewer could not trace, how long the grid took to assemble by hand. Tell us that and we will tell you plainly whether this fits.

Talk to usSee pricing

Get in touch

Talk to the people building it

No chatbot and no ticket queue. Tell us what your practice actually looks like — how many valuers, which asset classes, who reviews, which banks you are on the panel of — and someone who works on the software will reply.

info@legosphere.com

Please keep client names, owner names, survey numbers and exact addresses out of this box — it is an ordinary enquiry form, not a channel for a live mandate.