Uwais Iqbal, Founder of Simplexico AI, has spent years watching law firms get excited about AI, trial a few tools, and then quietly shelve them when nothing really changes. Despite two to three years of noise around generative AI, many firms are still not seeing real return on investment. Vendors promise transformation; lawyers see experiments, pilots, and “tinkering” – but little measurable impact.

Iqbal argued that this disconnect is less about the technology and more about the approach: most firms lack a structured methodology for identifying, designing, and deploying AI use cases. Instead, they buy shiny tools, bolt them on, and hope for magic.

A Seven-Step Framework for AI Adoption

Drawing on a decade of building AI in legal (from Thomson Reuters’ innovation lab to bespoke client projects), Iqbal outlined a practical seven-step approach that Simplexico AI now uses with law firms and in-house teams:

1. Education
   Build shared literacy first: what AI can and cannot do, its limitations, risks, and the genuine universe of possibilities. Without a common vocabulary, expectations quickly become unrealistic.
2. Discovery
   Map workflows, pain points, and opportunities across the organisation. Identify where AI could support strategy, not just “save time”. Prioritise and rank use cases based on impact and feasibility.
3. Use Case Design
   Take one high-value use case and design it properly. That means working with domain experts to clarify requirements, map the underlying workflow, and define how success will be measured.
4. Prototype or Pilot
   • On the buy side: run a focused pilot or POC with a vendor.
   • On the build side: develop a minimum viable solution to test with real users.
     The aim is evidence, not perfection.
5. Build or Procure at Scale
   Once validated, either go through a structured procurement process or build out a production-ready, integrated solution that fits your tech stack and security posture.
6. Change Management & Adoption
   Technology alone is not enough. Firms must plan how to onboard users, communicate the “why”, set expectations, and reduce friction so that AI tools actually make it into everyday workflows.
7. Ongoing Enablement
   The missing piece in many firms. AI roll-out is not a “go live and forget” exercise; teams need continuing support, training, examples, and refinement so that value compounds over time.

Although presented linearly, Iqbal stressed this is in reality an iterative loop, with feedback cycles between stages and constant refinement as teams learn.

Who Needs to Be Around the Table?

Traditional silos in law firms are a major blocker. Iqbal described the cross-functional group that tends to work best on AI projects:

• A technical lead (often an IT director)
• A partner sponsor in the relevant practice
• Innovation professionals (innovation lawyers / legal ops) to coordinate and project manage
• Practising lawyers and subject-matter experts as the source of workflow knowledge

This mix reflects a broader truth: integrating AI is not an IT project; it is an organisational change project that happens to involve technology.

Incentives, Billables, and Making Space for Innovation

Both Gannon and community members highlighted a familiar tension: time spent experimenting with AI is time not spent on billable work.

Iqbal pointed to emerging models where firms:

• Allocate explicit innovation time (for example, hours carved out for AI initiatives, as seen in leading firms experimenting with structured “AI time” for junior lawyers).
• Appoint innovation or AI champions with split remits between fee-earning and innovation.
• Introduce bonuses or other incentives for lawyers who lead successful AI initiatives.

The core message: if firms expect serious AI engagement, they must recognise and reward it – not treat it as something lawyers squeeze in after hours.

Build vs Buy: Where Is the Real Value?

The flood of legal tech vendors is not accidental. According to Iqbal, large language models now make it dramatically easier to build software, including by using AI to write much of the code. As a result:

• The model layer (OpenAI, Anthropic, etc., often accessed via Azure/AWS/GCP) is broadly available to vendors and firms alike.
• The application layer (the UI and basic software) is increasingly commoditised – and can often be replicated in weeks.
• The true differentiators are a firm’s data, workflows, and legal expertise.

This raises strategic questions:

• Should firms embed their most valuable workflows and knowledge into third-party tools?
• Who owns that IP if the relationship ends or the vendor is acquired?
• How easy is it to move to another tool without losing years of embedded know-how?

Iqbal’s view: firms should consider building where a use case is highly strategic and tied to their unique IP, and buying where tools are clearly commoditised and easy to replace.

Mid-Size and Smaller Firms: Underrated Advantage

While large firms have more budget and headcount, they are also more bureaucratic and, in Iqbal’s words, often more “wasteful” – pursuing AI for prestige and marketing rather than focused value.

Mid-size firms, by contrast, may be better placed to move quickly:

• They tend to have tighter, more coherent practice mixes (for example, a strong focus on real estate).
• They can laser in on a few high-impact workflows – such as lease review – and redesign them around AI, shifting from weekly manual exercises to same-day reviews with AI-assisted checking.
• With the right use case, Iqbal has seen mid-size firms identify seven-figure revenue opportunities simply by combining AI with new fee models (e.g., fixed-fee services powered by AI-enabled throughput).

For very small or solo practices, community member Michelle captured the concern: what if they simply ignore AI? Iqbal’s analogy was stark – refusing AI is like refusing the internet or electricity. Firms that choose “no AI at all” will, over time, struggle to remain competitive as clients and competitors move on.

The flip side: a solo practitioner who thoughtfully uses AI could genuinely compete with far larger teams by multiplying their personal capacity.

Education, Hesitation, and the Next Generation

Questions from in-house counsel and Indian legal ops professionals highlighted:

• The importance of mapping existing workflows before adopting tools – technology should follow clearly identified pain points, not lead them.
• Generational and cultural hesitation in some teams, where AI and “agile” are still unfamiliar concepts.
• A gap in law school education, particularly outside the UK/US, where students may graduate with little meaningful understanding of AI in practice.

Iqbal expects it may take up to a decade for the legal sector to achieve broad, baseline AI literacy. In the meantime, he advocates for practical, hands-on learning – working directly with tools, understanding prompts, workflows and model behaviour – rather than purely theoretical courses. Simplexico AI’s own “Legal AI Basics” email series is one example of efforts to close that gap.

Key Takeaway

• Start with education, not procurement. A shared understanding of AI’s capabilities, limits, and effort required is non-negotiable.
• Use a structured, seven-step approach. Jumping straight to tools without discovery and design almost guarantees poor ROI.
• Treat AI as organisational change. Cross-functional teams and proper incentives are essential for adoption.
• Protect your IP and workflows. Decide carefully what to build and what to buy, and be clear who owns the value you create.
• No firm is too small – or too big – to rethink its approach. Mid-size and even solo practices can use AI to punch well above their weight, but doing nothing is the riskiest choice of all.