Consultants Using AI Complete 12% More Tasks and Finish 25% Faster. GTA Firms Billing by the Hour Should Pay Attention.
Harvard and BCG tested 758 consultants. AI users finished 25% faster with 40% higher quality. Here is what that means for professional services in the GTA.
Harvard Business School and Boston Consulting Group ran one of the largest controlled experiments on AI and knowledge work ever conducted. They gave 758 BCG consultants access to GPT-4 and measured what happened. Consultants using AI completed 12.2% more tasks, finished 25.1% faster, and produced results that were 40% higher in quality compared to those working without AI. Consultants who previously scored below average improved their output by 43%. (Harvard Business School)
That study was published in 2023. Since then, McKinsey deployed its internal AI tool Lilli to over 45,000 employees — 72% now use it actively, averaging 17 queries per employee per week. McKinsey reports Lilli saves consultants 30% of their time on research and knowledge synthesis. (Future of Consulting)
BCG followed with Deckster, a tool that automatically reviews and polishes client presentations using best practices from senior leaders. About 40% of BCG associates use it weekly. (Future of Consulting)
The Big Four have noticed. Graduate hiring intakes across the UK branches of the Big Four dropped by 6% to 30% in the past two years, as firms determined that AI tools can handle many tasks previously assigned to junior analysts. (Future of Consulting)
This is not a future scenario. This is the operating reality at the firms GTA professional services companies compete against for talent, clients, and margin.
The Utilization Problem That AI Solves
Professional services firms live and die by billable utilization. The industry target is 70% to 80% — meaning seven or eight out of every ten working hours generate revenue. In practice, the global average has fallen to 68.9% and is trending downward. (Mosaic)
The gap between target and reality is not laziness. It is structural. Non-billable time — administrative tasks, proposal writing, internal meetings, business development, training, time entry — consumes 20% to 30% of a consultant's working hours. (MinuteDock)
For a 20-person consulting firm billing at $200 per hour, every percentage point of utilization represents roughly $400,000 in annual revenue. Moving from 69% to 75% utilization — a six-point improvement — represents $2.4 million in recovered billable capacity. The math is simple. The execution is where most firms stall.
The non-billable hours that drag utilization down follow predictable patterns. Proposal drafting. Client intake documentation. Meeting summaries. Status reporting. Research compilation. Knowledge retrieval from past projects. These are the tasks that AI handles well — not the strategic advisory work that justifies premium billing rates, but the administrative scaffolding around it.
Where AI Delivers Returns for Professional Services
Four areas consistently produce measurable results for GTA consulting, advisory, and accounting firms.
Proposal generation is the most immediate win. Teams using AI-assisted proposal tools report 60% to 70% reductions in first-draft creation time, transforming days of writing into hours of review and refinement. RFP requirement extraction — previously a 2- to 3-hour manual process per proposal — becomes near-instantaneous. AI handles 14 hours of work per RFP cycle that would otherwise consume senior billable time on non-billable activity. (Iris AI)
Client intake and document processing is the second area. Professional services firms running AI-powered intake workflows reduced processes that previously took up to four hours to under 30 minutes. Onboarding time dropped from 11 days to 3.5 days in one documented implementation, with compliance errors falling by 90%. (SmartDev)
Research and knowledge synthesis is the third. McKinsey's Lilli scans 100,000-plus internal documents in seconds and surfaces relevant precedents, frameworks, and data points that would take a human analyst hours to locate. For mid-market GTA firms without McKinsey's proprietary tools, autonomous AI agents can be configured to perform the same function across a firm's own document repository, CRM history, and project archives.
Internal operations represent the fourth category. A Grant Thornton study on Microsoft 365 Copilot deployment found an average of 7.5 hours saved per employee per week in knowledge-work contexts. (SmartDev) Osborne Clarke's AI-powered time capture pilot recovered 1.5 additional billable hours per user per week — time that was previously lost to incomplete tracking. (SmartDev)
Why GTA Firms Face This Sooner Than They Think
Toronto's professional services concentration is unusually high. The Toronto economic region accounts for 59.6% of Ontario's professional, scientific, and technical services workforce, with the sector representing 12.9% of total Toronto employment. (Job Bank Canada)
That density creates competitive pressure. When one firm in a market this concentrated adopts AI and improves utilization, client expectations shift. Response times compress. Proposal quality rises. Research depth increases. Competing firms do not get to choose whether AI changes their market — they only get to choose whether they adopt before or after their competitors.
AI adoption in Canada's professional, scientific, and technical services sector reached 13.7% in Q2 2024 — the second-highest adoption rate across all industries, behind only information and cultural industries. Another 26.6% of firms in the sector planned to adopt AI within the following 12 months. (Statistics Canada)
The firms moving fastest are not necessarily the largest. The Harvard/BCG experiment showed that below-average performers gained the most from AI — a 43% improvement compared to 17% for top performers. For mid-market GTA firms competing against Big Four offices with enterprise AI budgets, that finding matters. AI closes the capability gap from the bottom up.
What the Big Four Are Doing — and What It Means for Everyone Else
One analysis estimated that tools like McKinsey's Lilli and BCG's Deckster can now perform roughly 80% of a junior analyst's typical research and slide-generation work — and do it in seconds. (Future of Consulting)
This is not replacing consultants. It is changing what a consultant's hour is worth. When an associate at a Big Four firm spends two hours on research that AI could complete in five minutes, the firm is paying associate rates for work that produces near-zero marginal value. The firms that recognized this first cut junior hiring and redirected those roles toward AI-augmented workflows.
For GTA firms with 10 to 100 employees, the implication is different but equally significant. These firms cannot afford to maintain a bench of analysts for research and slide work. But they also cannot afford to have $300-per-hour partners spending 30% of their time on non-billable administrative tasks. AI agents handle the administrative layer — proposal first drafts, meeting summarization, client intake processing, internal knowledge retrieval — so that every human hour goes toward the advisory work clients actually pay for.
PwC's analysis of one billion job postings found that the AI wage premium jumped from 25% to 56% in a single year. Consultants who specialize in AI-augmented delivery now command 30% to 40% fee premiums over generalists. (Deltek) The market is pricing AI capability into professional services — not as a novelty, but as an expected standard.
What This Looks Like in Practice
A 25-person management consulting firm in Mississauga processes 40 proposals per quarter. Each proposal takes 8 to 12 hours of non-billable time across research, drafting, formatting, and internal review. That is 320 to 480 hours per quarter — the equivalent of two full-time employees doing nothing but proposals.
An AI agent configured for that firm's proposal workflow handles the first draft, extracts requirements from RFPs, pulls relevant case studies from the firm's project archive, and formats the output to match brand standards. The human role shifts from writing to reviewing and approving. Proposal cycle time drops from days to hours. Those 320 to 480 hours convert back to billable capacity.
Apply the same logic to client intake documentation, meeting transcription and summarization, competitive research, and internal knowledge management, and the utilization impact compounds. A firm that recovers even five billable hours per consultant per week at $200 per hour adds $1,000 per consultant per week in recovered capacity — $52,000 per consultant per year.
DeployLabs builds autonomous AI agents for professional services firms across the GTA. Not off-the-shelf software that requires your team to learn another platform — coordinated AI systems that integrate with your existing tools, operate on your data, and execute the repetitive workflows that are dragging your utilization below target. The engagement starts with a $2,500 AI readiness assessment — a structured evaluation of where AI delivers the highest return for your specific operation, credited in full toward any build. Custom systems start at $7,500, with ongoing support at $2,000 to $5,000 per month.
Book a discovery call to see where your firm's non-billable hours are hiding recoverable revenue.