The Relationship Graph as a Firm Asset (Not a Database)

A database answers questions about objects. A graph answers questions about connections. Most law-firm CRMs are databases pretending to be graphs, and that gap is the reason partners stop running queries they cannot get useful answers to.

The reframing matters because the firm's commercial value sits in the connections, not the rows.

Database thinking versus graph thinking

Database thinking asks: who is at Acme right now, what is their email, when did we last contact them. These are useful questions. They are also the easy ones.

Graph thinking asks different questions. Who in the firm has the strongest path to the new general counsel at Acme — through which prior matter, which mutual contact, which conference. Which referrer has sent us five files over three years that we have never thanked properly. Which inventor named on a competitor's recent filing is two hops from a partner we already know.

A database can be coerced into answering some of these with enough joins. A graph answers them natively, and the firm starts asking questions it would not have bothered to ask before.

Why IP firms benefit disproportionately

Most professional-services firms have a roughly linear relationship structure: in-house buyer, outside counsel, done. IP firms do not. The relationship structure of an IP practice is dense and indirect:

• Foreign associates who refer prosecution work and never appear on a matter for the client directly
• In-house IP counsel who move between operating companies every three to five years, carrying the relationship with them
• Inventors named on filings who later become founders, who later become clients
• Patent attorneys at competitor firms who refer conflicts work
• Litigation co-counsel who become referrers on the next case
• Mutual contacts at standards bodies, industry associations, university tech-transfer offices

A database that treats each of these as an isolated contact loses the structure. A graph keeps it. We wrote separately on why the indirect-client network is the under-managed half of an IP firm's BD, and graph thinking is the foundation of that argument.

What the graph changes about firm queries

Once the firm models its relationships as a graph rather than a table, three categories of question become tractable.

Pathing. Given a target organisation, what are the shortest and strongest paths from the firm to a decision-maker there. The answer might run through a former colleague of a senior associate rather than through a partner.

Centrality. Which contacts sit at the centre of the firm's network — touching many matters, many partners, many referrers. These are the relationships that, if lost, would cost disproportionately. They are also the ones most worth investing in.

Cluster detection. Which groups of contacts move together — the in-house team that hires the same outside counsel, the inventor cohort that spins out into a wave of startups, the litigation circuit that recycles co-counsel. Spotting clusters early is the difference between catching a market wave and reading about it in the legal press.

None of these queries are exotic. They are obvious once the data model supports them. They are impossible when it does not.

What this means for the system the firm buys

If the firm accepts that its relationship structure is a graph, the implication for the CRM is concrete: the system has to model relationships between people, between people and organisations, between people and matters, between organisations and organisations, with first-class entities for each. A bolt-on "relationship strength" field on a contact record is not a graph. It is a database trying to look like one.

The firms that get this right stop treating their CRM as a system of record and start treating it as a system of inference. That is the threshold worth crossing.

Related: The Indirect-Client Network: The Hidden Half of IP Referrals. The CRM as the Operating System of a Modern Law Firm. The Best CRM for IP Law Firms in 2026.