At TELUS, the work started with a concrete problem: a complex portfolio of contracts and a legal team that needed real insight, not just faster summaries. Instead of throwing more junior lawyers at the work, Rahim Esmail stayed up late developing an AI-powered software application designed to apply senior-level legal reasoning at scale. He then put that tool, now known as the Document Analyzer, into his colleagues’ hands.
Esmail, who is associate general counsel and co-chair of TELUS’s AI Think Tank, was “vibe coding,” using a new wave of AI tools to build software without writing traditional code. The process uses natural-language instructions and AI coding tools to assemble working software without a computer science background, enabling people with no coding experience to design applications tailored to their specific needs. Because these tools have removed barriers to software development, Esmail predicts that early adopters in the legal industry will start moving away from one-size-fits-all, expensive SaaS platforms toward highly specific, custom-developed tools.
“On a one-off basis, LLMs are at the point where a user could upload a document, ask a question, and get the output needed,” says Esmail, but that still meant prompts, re-prompts and manual checks, all on a small batch basis.
The Document Analyzer is his answer to that bottleneck. The app ingests large folders of unstructured files, strips out the text and feeds it into a large language model using company standards and prompts tuned to the issues that matter to TELUS. It then compiles the answers into a table that shows relationships between documents, flags issues, and sets out the key fields the business needs to track. The tool also rolls those patterns up into “a two-page executive summary” of frequencies, risks and themes, he says. A legal team that once had to choose between superficial sampling and unaffordable full-scale review can now see the whole landscape and decide where to act.
In Esmail’s view, “scale and efficiencies are table stakes,” and the real competitive edge comes when lawyers blend their expertise with technology to raise the standard of legal advice. Esmail draws a parallel to quantitative trading firms. The goal isn't just to execute faster, but to increase sophistication by uncovering macro-level risks and patterns that traditional legal review cannot identify. In a classic M&A review, the Document Analyzer can sweep thousands of agreements for the exact issues the client cares about, from change of control clauses to revenue protections, and lay them out in a format business leaders can scan in minutes. In a live book of customer contracts, the same engine can surface restrictive covenants, end of life terms related to legacy services, or pricing levers that allow a company to rebase revenue. Ultimately, Esmail says, “this transforms contracts into a dataset that can be used to provide strategic legal advice and business intelligence."
Under the hood, the app depends on capabilities that Esmail believes many lawyers still underestimate. He argues that modern language models are strong at extracting information from contracts and now handle the messy internal logic that slows traditional review work. “We are way past simple text extraction,” he says. “These models can now execute high-fidelity legal reasoning, including weighing conflicting terms and rationalizing deep cross-references to synthesize an accurate legal conclusion.”
Those reasoning skills are embedded in standard analysis types he has pre-built into the tool. Users select from a dropdown menu for recurring tasks, and the prompts sit behind the interface so lawyers and contract managers are not forced to become prompt engineers. The more ambitious feature is a custom analysis mode. A user describes a goal in plain language, the app proposes fields to extract, the user edits them, and the system then runs a pass across the entire document set. That mode turns the Document Analyzer into what Esmail calls “a dynamic intelligence engine for any type of analysis.” Demand within TELUS is coming not only from lawyers but also from procurement and sales teams that want the same visibility into their agreements, he says.
For now, the prototype he built is being scaled up for enterprise use. Internal developers are rebuilding it to integrate with TELUS systems and scale across business units. TELUS keeps “a human in the loop,” i.e. a lawyer, for legal analysis and has been conducting “robust checks” as the tool rolls out, he says.
The Document Analyzer grew out of a longer arc that pulled Esmail steadily closer to the technology side of the business. Called to the bar in 2011, he started at Borden Ladner Gervais in Vancouver, doing broad corporate-commercial work, before moving to McCarthy Tétrault to focus on technology-heavy transactions. A secondment to BMO’s technology procurement group gave him his first sustained exposure to working with a single client on repeat mandates. It showed him how legal advice changes when a lawyer understands the business over time rather than dropping in on isolated files.
After he joined TELUS, he spent his first year on the sales side working on “large complex deals” for governments and enterprises. He then shifted to the procurement team, where he was responsible for “supporting TELUS’ digital transformation, including its pioneering shift to cloud infrastructure.”
Today, in his role at TELUS Agriculture & Consumer Goods, he sits where those strands converge. In his words, this division of TELUS is largely “a SaaS and data business” that helps crop growers optimize inputs, supports consumer businesses with trade promotion analytics, and works with veterinarians to improve outcomes in animal agriculture. The business runs on software and data, and Esmail wants the legal function to operate on the same footing.
He argues that the newest models are already capable of work that used to require large teams, and he says legal leaders must now decide whether to use that capability only to cut time from existing tasks or to change how legal services are delivered fundamentally. His own bet is firmly on the latter. He wants legal teams to combine “legal skills with custom code and an LLM,” insisting that “those three things together” can generate work product at a level that he or even a full traditional team could not do alone. Ultimately, Esmail believes that technological evolution has created an opportunity for legal teams to serve as architects of proprietary tools that can scale their expertise across an entire organization.
He notes a wider community of lawyers experimenting with similar techniques, from global WhatsApp groups of vibe coding practitioners to self-described quant lawyers. Many are still working quietly inside cautious organizations but are moving along the same direction, even if its endpoint is not yet clear.