DataFab //  artificial proprietary intelligence T+00:00:00 ISO/IEC 27001:2022 · SOC 2 Type II — certified Raw egress  0 B 

DataFab  /  Utilities  /  Client lifecycle management

Governed utility · legal & professional services

One resolved client, from prospect to exit.

Client lifecycle management over the systems a firm already runs — matter management, the document system, the practice and finance estate, and the CRM. Parties, terms and dates extracted from the engagement the firm actually signed; obligations tracked with the clause they came from; conflicts checked against the resolved graph rather than a name index.

Client Life Cycle ManagementConflicts on the graphDataFab does not train on client data
What a firm actually holds

“Everything we know about a client is spread across matter management, the document system, the finance ledger and the heads of the people who did the work — and none of it survives the person leaving.”

A firm’s knowledge is its only asset that does not appear on the balance sheet, and it is the one asset it manages worst. Not because the systems are bad, but because no question can cross them: the client in the CRM, the client in the matter system and the client in the ledger are three records nobody has agreed are one.

Resolve them and the lifecycle becomes a loop rather than a set of handoffs. Conflicts are checked against what the firm knows rather than what somebody typed. Obligations carry the clause they came from. And the expertise applied a hundred times a week becomes something the firm owns rather than something it rents from whoever happens to still be there.

The lifecycle

A loop, not a funnel.

Prospect, conflicts, onboarding, engagement, matter, obligations, billing, renewal or exit — each writing back to the same resolved client rather than into its own system of record.

Fig. 01  //  the client lifecycle, on one resolved client◆ every stage writes back
THE CLIENT LIFE CYCLE IS A LOOP, NOT A FUNNEL. EVERY STAGE WRITES BACK TO THE SAME RESOLVED CLIENT.ProspectConflictsOnboardEngageMatterObligationsBillRenew / exitOne resolvedclientacross matter, document, finance and CRMConflictsChecked against the resolved graph, not a name index — and re-checked as it changes.Engagement termsParties, scope, rates and obligations extracted from the letter you actually signed.ObligationsTracked with the clause they came from, alerted before they bite, not after.Matter intelligenceWhat the firm already knows about this client, this counterparty, this clause.Ethical wallsDeclared over entity and relationship types, enforced as the graph is traversed.Renewal & exitRetention, destruction and disengagement obligations surfaced on schedule.DATAFAB DOES NOT TRAIN ON CLIENT DATA · PRIVILEGED MATERIAL IS RESOLVED IN PLACE, UNDER THE FIRM’S OWN CONTROLS

Conflicts

The conflict nobody wrote down.

A name-index search finds the conflicts somebody already knew about. The dangerous ones are the conclusions that exist only between records: a directorship in a filing, a shareholding in a register, a matter in the document system — each innocuous alone, and together putting the firm on both sides.

Checked against the resolved graph, the search runs on the real party in every rendering of its name, across every system the firm runs, and returns the connection with the evidence for it. And because the graph is live, the check is repeated when the graph changes rather than only when somebody opens a new matter.

Resolved

The party, in every rendering

Aliases, legal names, transliterations and former names collapse to one entity before the check runs.

Relational

One and two hops out

Group companies, common directors, shared beneficial owners and counterparties already adverse in another matter.

Continuous

Re-checked, not re-run

A new filing or a new matter reopens the check automatically, and the firm is told what changed rather than handed the whole search again.

Evidenced

The clearance is a record

What was searched, what was found, what was cleared and on whose authority — produced as the work happens.

Terms & obligations

Extracted from the engagement you actually signed.

Parties, scope, rates, caps, notice periods, retention terms and reporting duties read out of the engagement letter and its variations, bound to schema, and tracked as live obligations with the clause they came from attached.

Extraction

Parties, terms, dates

Per-field confidence on every extracted term, with the page and passage it came from — and a flag rather than a guess where the wording is ambiguous.

Tracking

Alerted before they bite

Reporting duties, notice periods, rate reviews and caps surfaced ahead of the date rather than discovered after it.

Review

Clause level, governed

Deviation from the firm’s standard position identified clause by clause, with the standard it deviates from shown alongside.

Lifecycle

Intake through renewal

Retention, destruction and disengagement obligations surfaced on schedule at the end of the relationship, not left to a diary entry.

Ethical walls

The wall has to hold at the ontology.

Once a firm’s estate is resolved into a graph, blocking a matter file no longer blocks the conclusion — because the conclusion was distributed across a billing entry, a calendar, a conflicts search and a time narrative, each harmless alone. The barrier has to be declared over entity and relationship types, and enforced as the graph is traversed.

Fig. 02  //  document-level wall versus ontology-level wall◆ enforced at traversal, before planning
A DOCUMENT WALL ASKS WHETHER YOU MAY OPEN SOMETHING. AN ONTOLOGY WALL ASKS WHETHER YOU MAY RECONSTRUCT WHAT IT WOULD HAVE TOLD YOU.DOCUMENT-LEVEL WALLThe file is closed. The fact is not.Matter file — blockedBilling entrya matter exists for this counterpartyCalendartwo partners met them on TuesdayConflicts searchsomebody already checked this nameTime narrativethe subject of the work is inferableTHE WALL HOLDS. THE CONCLUSION LEAKS ANYWAY.ONTOLOGY-LEVEL WALLGoverned at the type, not the fileEntity typeswhich classes of entity a role may see at allRelationship typeswhich edges may be traversed, and in which directionAttribute policywhich properties resolve, which are masked at readInference suppressionderived edges withheld where the premise is walledAggregation limitscounts and patterns that would reveal a walled factTraversal-time enforcementreasoning runs over a view that already excludes itTHE ANSWER CANNOT BE REBUILT FROM THE PIECES THAT WERE LEFT OPEN.INHERITED ENTITLEMENTS ARE THE FLOOR — STRICTER, NEVER LOOSER · EVERY REFUSAL IS RECORDED, INCLUDING THE ONES A USER NEVER SEES

Matter intelligence

Be the platform, not the reseller.

The judgement a firm applies a hundred times a week currently lives in people and leaves with them. Encoded once in the Knowledge & Agentic Studio it becomes a governed, versioned, reusable object the firm owns outright — deployable across every team, every client and every jurisdiction, without a vendor in the loop.

Precedent that finds you

A change in a client’s world, resolved against what the firm already knows, delivered as a brief rather than a search result.

What we already know

What the firm has advised this client, this counterparty and on this clause — across matters, without breaching a wall.

Ask the matter

Any matter questioned in plain language across the connected systems, answered with the evidence behind it.

The firm’s own agencies

Composed by the people who do the work, from documents the firm already has, and reused by everyone entitled to them.

A standing commitment

DataFab does not train on client data. Client work product, matter content and privileged material are never used to train models — the value comes from resolving what the firm already holds, in place, under the firm’s own controls.

Next step

Bring the conflicts check that takes three days.

Client lifecycle management over the systems the firm already runs, with the conflicts search as the first proof.