Our Take
Allgrove is right about vendor lock-in risk, but his warning about pricing pressure is the real story: as compute subsidies end, law firms will face a squeeze between encouraging AI use and paying consumption-based bills.
Why it matters
Large law firms are making AI strategy decisions now that will stick for years. Allgrove's multi-vendor stance challenges the industry's assumption that a single legal AI platform can solve everything, which matters as pricing models shift from flat to per-use.
Do this week
General counsel and legal operations: audit your current vendor contracts for auto-renewal clauses and exit costs before signing any multi-year AI platform agreements.
Baker McKenzie's multi-vendor, multi-model strategy
Ben Allgrove, chief innovation officer at global firm Baker McKenzie, outlined a deliberate strategy to avoid locking the firm into a single AI vendor or model. The rationale is simple: the pace of change in generative AI outpaces any firm's ability to predict strategic needs beyond six months.
Rather than picking one vendor and consolidating, Baker McKenzie is working with multiple vendors and models while standardising tools within specific practice areas or use cases. This prevents end users from facing overwhelming choice while preserving the firm's flexibility to swap tools as the market evolves.
Allgrove emphasised that multi-vendor does not mean "anything goes." Training focuses on core concepts and foundational technology understanding rather than tool-specific instruction. This reflects a shift in what high-performing lawyers now need: continuous learning and adaptability.
Control of the orchestration layer is the real lever
The firm's strategic focus is retaining control of its orchestration layer—the layer that sits above individual AI tools and coordinates them—rather than handing that control to an external platform vendor. This matters for cost management, intellectual property protection, client relationships, and scalability.
Allgrove was explicit about a tension emerging in the market: legal-specific AI vendors must justify their value as general-purpose models from companies like Anthropic improve. The differentiator between law firms will not be the technology itself. Instead, human talent and proprietary knowledge will remain the primary source of competitive advantage. Technology shapes how services are delivered and priced, but it does not itself define advantage.
The unspoken second-order effect: if technology is not the differentiator, firms that lock into expensive single-vendor platforms will have less margin to invest in talent and knowledge.
Watch the pricing shift, not the model releases
Allgrove flagged a critical economic inflection point. Current pricing dynamics are supported by subsidised compute costs. As those subsidies end and consumption-based pricing takes hold, law firms will face a paradox: they need to encourage AI adoption to realise its efficiency gains, but higher usage directly increases costs. This creates pressure on traditional time-based billing models.
This is where the multi-vendor strategy has teeth. Firms locked into a single vendor's consumption-based pricing will have limited negotiating power or alternatives. Firms with multiple vendors and direct control over orchestration can shop for capacity, negotiate discounts, and shift workloads if pricing becomes uncompetitive.
Allgrove acknowledged this AI wave is different from previous hype cycles: its general-purpose nature and cross-industry impact make it materially different. But he resisted the marketing narrative, noting that differentiator between firms will not emerge from choosing the "right" model.