For consulting engineering firms

AI for engineering firms: more capacity, same seal

Your engineers already use Copilot or ChatGPT, your board wants a plan, and nobody wants AI anywhere near a stamped drawing without a clear rule. This page shows where AI earns its keep in a consulting engineering firm, what to keep a licensed engineer on, and how to turn scattered pilots into hours you can actually measure.

Start with the work that never needs a seal: pursuits, proposals, project history, QA lookups.
Keep the engineer of record in control, with checks your quality manager will sign off on.
Measure hours returned and turnaround times, not logins.

Book a strategy call Have our engineers review your idea →

The landscape

Where AI stands in engineering firms

Engineering firms are not behind on AI. Statistics Canada's 2026 business survey puts professional, scientific and technical services among the three industries most likely to use AI, at close to one in three businesses. Most firms you compete with have licences, a policy in draft and a handful of keen early users.

The use is lopsided, though. Firms are comfortable using AI to write marketing copy, summarize meetings and draft emails. They are far more careful near design and delivery, and they should be. ACEC-Canada's 2026 sentiment survey says the same thing plainly: firms are adding AI to business processes, administration and marketing, while use in core engineering and design work remains limited.

The bigger gap is between using AI and getting something from it. Deltek's 2026 Clarity study of nearly 900 North American A&E firms found AI adoption jumped to 70%, but only 38% reported a measurable positive business impact. Licences went out. Workflows mostly stayed the same.

Meanwhile the pressure that makes AI worth doing hasn't gone anywhere. ACEC-Canada members report open positions at 81% of firms, the federal government projects a moderate risk of shortage for civil engineers through 2035, and every retirement takes years of project history with it. The firms that pull ahead will be the ones that give their engineers back hours, not the ones with the most tools.

32.4%

of Canadian professional, scientific and technical services businesses used AI to deliver services in the past year, versus 19.2% of all businesses

Statistics Canada, 2026 ↗

70% vs 38%

of A&E firms now use AI, but only 38% report a measurable positive business impact from it

Deltek Clarity A&E Study, 2026 ↗

81% vs 36%

of engineering firms use AI in marketing and sales, but only 36% in project design and delivery

ACEC Research Institute (BST Global AI + Data Survey), 2025 ↗

81%

of Canadian consulting engineering firms have at least one open position

ACEC-Canada Engineering Business Sentiment, H1 2026 ↗

Why generic advice fails

What makes engineering firms different

Most AI advice is written for sales teams and call centres. It skips the things that decide whether AI works in a firm that seals drawings and bills by the hour.

The seal doesn't move

In Ontario, PEO has adopted Engineers and Geoscientists BC's practice advisory on AI as a guideline. The core message: you remain professionally responsible for work that includes AI output, and you shouldn't rely on AI for safety or environmental work unless you understand how it got its answer. Engineers Canada's national position says the same. So the question is never "can AI do this?" It's "can a licensed engineer check this, and will they want to?"

You sell hours

If a report takes half the time on an hourly contract, revenue drops unless the firm wins more work or prices differently. ACEC-Canada's white paper flags remuneration directly, and the ACEC Research Institute names time-on-task billing as a real barrier. Point AI first at unbilled work (pursuits, proposals, admin) and the economics are simple. Point it at billable work and you need a pricing conversation too.

Your know-how lives in old projects and senior heads

The answer to most questions in an engineering firm already exists: in a 2017 report, a past proposal, a calc package, or a principal's memory. That's good news, because AI is strong at finding and reusing what you already wrote. It's also urgent, because those people retire and the folders get harder to search every year.

Client data comes with strings

Utility drawings, municipal infrastructure data, forensic claim files and NDAs all limit where information can go. ACEC-Canada's guidance is blunt: never share confidential client data with public tools. That rules out a lot of free-tool experimentation and makes your own environment, with proper permissions, the default.

QA/QC is already your culture

Engineers check things for a living. The advisory PEO adopted expects AI use to fit documented checks: note the tool and version, test it on known cases, keep inputs and outputs where the result matters. Firms that build AI into the existing review process get adoption. Firms that bolt it on beside the process get quiet resistance.

Where to start

Where AI pays off first

Start where the work is high-volume, repetitive and checkable, and where nobody has to stamp the output. Each one comes from a build or coaching session we've delivered, and the same pattern fits consulting engineering. Browse the full list in the Use Case Explorer.

VP Business Development

Every RFP you're qualified for, ranked each morning

Municipal, provincial and private bid sites get checked when someone has time, so good-fit RFPs get found with four days left. A tool that watches every source and scores each opportunity against your disciplines and past wins gives your go/no-go meeting one list instead of a dozen tabs. It costs no billable hours and touches no sealed work.

A build we’ve deliveredFinding the opportunities you were missingBid sites watched now and then, to all the time · About 3 months

See the build →

Marketing and proposal manager

Proposals, résumés and project sheets drafted inside Microsoft 365

Your marketing team rebuilds the same qualifications, project descriptions and CVs for every submission. Drafting from your own past proposals, with people search over who did what, cuts the hunt and keeps one voice across offices. Because it lives in the apps your team already has open, there's no new tool to learn.

A build we’ve deliveredProposal writing inside the tools you already useHundreds of hours a year of archive search returned · First release, then four extensions

See the build →

Principals and BD leads

Which past projects, and which PM and lead, fit this RFP?

"Have we done a bridge rehab with this owner?" usually means asking around the office. Matching the RFP's evaluation criteria to your completed project records and staff history surfaces the right references and the teams that have delivered together. Pursuit decisions rest on the record, not on who happens to remember.

A build we’ve deliveredWhich of our past projects fits this bid?Asking around the office, to a search · About 3 months

See the build →

Principals and project controls

Fee estimates and level of effort from your own history

Senior engineers spend hours building hours-by-task fee sheets by copying the last similar job. Predicting the hours from past projects, with a confidence rating on each line, leaves them to judge the scope instead of retyping it. Fees become more consistent, and you can show how you got the number.

A build we’ve deliveredEstimates without starting from a blank spreadsheet15+ senior hours a week returned · About 4 months

See the build →

Discipline leads and QA/QC

Submittals and shop drawings checked against the design

MEP and structural teams compare vendor submittals to drawings and equipment schedules sheet by sheet. An agent loaded with your own review guide lists every mismatch for the engineer to confirm, so the engineer reviews findings instead of hunting for them. The engineer still decides; the tool does the first pass.

A build we’ve deliveredDrawings checked against vendor submittalsSheet-by-sheet comparison, to a discrepancy list · Built in a coaching session, shared with the team

See the build →

Quality manager

Your QMS and standard procedures, answered with the section cited

Staff search the quality manual and standard procedures at the worst moment: mid-task, with a deadline. An assistant limited to your approved documents gives the working answer and cites the section, and refuses anything outside that set. Juniors stop interrupting seniors, and everyone works from the same controlled source.

A build we’ve deliveredStep-by-step procedures, cited to the sectionProcedure lookups without stopping work · Weeks, on the existing foundation

See the build →

Engineers and analysts

Public data pulled into your analysis on request

Engineers download weather records, climate normals or other public datasets by hand from awkward government sites before an analysis. This is common in forensic, environmental and stormwater work. An agent that fetches the dataset and returns a clean file saves that time on every job. It was built live in a coaching session.

A build we’ve deliveredPublic data pulled into your model on demandManual downloads, to a clean file on request · Built live in minutes

See the build →

Department heads and operations

Senior know-how captured before it walks out the door

The process for a recurring report or calculation often lives in one person's head. Record a call while they narrate it on screen, and Copilot drafts the procedure against your template or sets up a working agent. It's a cheap, practical start on succession.

A build we’ve deliveredRecord a process once, get an SOP or an agentOne recording, to an SOP or a working agent · First version in an afternoon

See the build →

Browse all the builds in the Use Case Explorer →

The people side

Why AI adoption stalls in engineering firms

When AI stalls in an engineering firm, the cause is rarely the software. It's almost always people, time and rules.

1Nobody knows what is safe, so nobody tries

Engineers are trained to avoid risk to their licence. Without a clear line between "use freely" (proposals, minutes, research) and "check before relying on it" (anything near design), the careful majority does nothing while a few use free tools quietly. A short, specific AI policy written in your firm's language, plus six weeks of practice through an AI Habit Builder, does more than another licence purchase.

2Utilization targets leave no room to learn

When people are measured on billable hours, an hour spent learning a tool looks like a cost. AI-First Moments workshops fix this by having teams bring a real workflow, like a proposal section or a site review report, and leave with it done faster, so the learning pays for itself the same day.

3Seniors set the tone

ACEC's research lists cultural resistance among senior and highly specialized staff as a top barrier. If principals don't use it, juniors take the hint. Six weeks of leadership sponsorship, where leaders use AI on their own work and say so out loud, changes the signal faster than any memo.

4Success is measured in logins

Licence counts and active users tell you nothing about whether proposals go out faster or seniors get time back. An AI Adoption Diagnostic scores people capability, work design, AI's role and enterprise adoption, then names the one constraint holding you back. From then on you track cycle time and capacity created.

How our AI Transformation & Adoption practice works →

Your first year

A 12-month plan for the executive leading AI

This sequence works for a firm of 100 to 5,000 people with licences already out and uneven use. The details change by firm; the order rarely does. It follows our Three-Layer AI Playbook.

First 30 days

Get honest about where you are

Run an AI Adoption Diagnostic to see who uses AI, for what, and which single constraint is holding the firm back.
Publish a one-page AI rule set tied to PEO guidance: what's fine, what needs a documented check, what's off-limits.
Fix Microsoft 365 permissions on project and HR sites before you widen Copilot access.
Pick two baseline numbers to beat, such as proposal turnaround and hours to build a fee estimate.

Days 30 to 90

Win where no seal is involved

Run an Immersion Day and the Executive AI Program so principals lead by using it, not delegating it.
Hold AI-First Moments workshops with BD, marketing and one delivery team on their own workflows.
Build a ranked use-case backlog, scored on hours returned, risk and how checkable the output is.
Start the first build on pursuits or proposals, in your own Azure tenant.

Months 3 to 6

Build habits and the first custom tool

Launch an AI Ambassador network with one person per office or discipline.
Run a six-week AI Habit Builder for project managers and senior engineers.
Ship the first custom build and track usage by team from day one.
Add a QA lookup or submittal-review agent with documented checks your quality manager signs off.

Months 6 to 12

Scale what's working, retire what isn't

Report outcomes to the board: proposals out, turnaround time, senior hours returned.
Extend proven tools to more offices and disciplines before starting new ones.
Decide where AI touches billable work and update pricing and client agreements to match.
Refresh the backlog and re-run the diagnostic to confirm the constraint has moved.

Guardrails

Risk, data and governance

Governance in an engineering firm isn't a separate project. It sits on top of the quality system you already run. Four areas need attention first.

Professional responsibility and documented checks

The advisory PEO adopted is clear that responsibility rests with the engineer whether or not AI was used. Where AI generates results or informs design decisions, document its use as if the engineer did the research or calcs: tool and version, inputs, outputs, and the validation run. Keep AI for drafting and finding, and keep the engineering judgment with a licensed person.

Client confidentiality and Copilot oversharing

Copilot answers from anything a user already has permission to see. In most firms that includes project folders shared too widely years ago. Microsoft's own deployment guidance starts with fixing oversharing before rollout. Where teams serve competing clients, build walls into the tool so one team's search never reaches another team's files.

Contracts, insurance and disclosure

ACEC-Canada's white paper reports that professional liability policies currently carry no AI exclusion, but insurers expect clear scope and documented decisions; confirm with your own broker. Its white paper also says AI use should be disclosed to clients. Review your standard agreements, and expect public clients to ask: Ontario's Bill 194 lets the province set AI accountability and risk rules for ministries, municipalities and other public sector entities.

Privacy law and personal information

Forensic claim files, site photos, HR records and résumés all hold personal information covered by PIPEDA or provincial law. Canada's privacy commissioners have said generative AI sits inside existing privacy law, not outside it. Keep personal data in your own environment, limit what each tool can reach, and know where your data is processed.

Client stories

Firms that bid and deliver like you

These are leaders from engineering, testing and construction firms who started with one problem and kept going.

“ForwardPath has delivered AI tools that our teams use every day. More departments keep adopting these tools, and ForwardPath has been a reliable, long-term partner every step of the way.”

Phil LangilleVP, Business Development, Rimkus

An engineering and consulting firm that automated proposal generation and tender discovery, with more departments adopting the tools over time.

“We immediately started seeing quick wins with the AI Tool. It has really increased our momentum and created a lot of flow.”

Julie WestSVP, Business Improvements & Compliance, Kinectrics

A testing and engineering firm in energy and utilities that started with a company knowledge tool and saw quick wins and growing AI literacy across leadership.

“Our first build changed how we pursue work. Our proposal content, resumes, and project history now live in one place. Our team can research, draft and build submissions much faster than before, with better consistency. ForwardPath took the time to understand how our business actually bids, and we’re already looking at what’s next with them to find and qualify tenders earlier.”

Jerry RyndaVP, Business Development & Marketing, Buttcon

A contractor that pursues work the way consulting firms do, now drafting submissions from one home for proposal content, résumés and project history.

Read all our client stories →

Before you spend

Questions to ask before you spend another dollar on AI

Ask these of your own team and of any partner pitching you. Vague answers are a warning sign.

Which of our hours are unbilled and repetitive?

Pursuits, proposals, fee estimates and admin are the safest place to start, and every hour saved there goes straight to margin.

Where would an engineer have to check the output, and would they?

If the check takes as long as doing the work, it's the wrong use case. Pick work where reviewing is much faster than producing.

Who can see what in our Microsoft 365 tenant today?

Copilot exposes existing permission mistakes. You want to find them before a junior asks it about salaries or another client's project.

What happens to our revenue on hourly contracts if delivery gets faster?

Decide early whether savings fund more wins, better margins or new pricing, or the CFO will slow everything down later.

Where does our data go, and who owns the code?

Builds should run in your own cloud, under your contracts, with code you own, so you're not renting your own know-how back.

How will we know it worked in six months?

Name the number now: proposal turnaround, submittal review time, senior hours returned. Logins don't count.

What will our principals do differently on Monday?

If leaders don't use AI on their own work, adoption below them stays shallow, whatever the training budget.

FAQ

Frequently asked questions

Can engineers use ChatGPT or Copilot on project work?

Yes, with limits. Engineering regulators haven't banned it, but the guidance PEO adopted says you stay professionally responsible for anything that includes AI output. Use it freely for drafting, summarizing and finding information. Where it informs a result or design decision, record the tool and version, validate the output and keep the records. Never paste confidential client data into a public tool; use an enterprise version inside your own Microsoft 365 or Azure environment.

Does PEO have rules on AI for engineers in Ontario?

PEO has adopted Engineers and Geoscientists BC's practice advisory, Use of Artificial Intelligence in Professional Practice, as a guideline in Ontario, read alongside Ontario's statutes and PEO's Code of Ethics. It covers risk assessment, direct supervision, documented checks, record retention and independent review when AI is used. The key point: responsibility for the work product rests entirely with the engineer, regardless of whether AI tools were used.

Where should an engineering firm start with AI?

Start where no seal is involved and the work repeats: finding tenders, drafting proposals, matching past projects to an RFP, building fee estimates and answering QMS questions. These save senior time without touching design liability, and they build the habits and governance you'll need later. Pick one or two, set a baseline number, ship a working tool in your own environment, then expand to work closer to delivery.

Will AI reduce our fees if we bill hourly?

It can, if you speed up billable work and change nothing else. ACEC-Canada's AI white paper notes that billing approaches may need to change as AI shortens tasks. Most firms start with unbilled work like pursuits and proposals, where every hour saved goes to margin or more bids. For billable work, decide up front whether savings fund more capacity, better margins or fixed-fee pricing.

Is Microsoft Copilot safe for confidential client files?

Copilot answers using content each person already has permission to access, so it is only as safe as your permissions. Many firms have project and HR sites shared far more widely than anyone intended. Microsoft's own deployment guidance starts with finding and fixing oversharing, then setting guardrails. Do that work before a wide rollout, and use separate access controls where teams work for competing clients.

How do we measure whether AI is working in our firm?

Measure outcomes, not usage. Pick numbers your leadership already cares about: proposal turnaround time, bids submitted per BD person, hours to build a fee estimate, submittal review time, senior hours returned to billable work. Set a baseline before you start, and track usage by team so you can see which tools landed. Deltek's 2026 study found only 38% of A&E firms report measurable impact. A baseline is how you prove you're one of them.

Start the conversation

Lead AI in your engineering firm with confidence.

Start with a conversation with our leadership team. We’ll look at where you are, where AI can move the needle first, and what your next 90 days could look like.

Book a strategy call Have our engineers review your idea →

Sources

EGBC Practice Advisory: Use of Artificial Intelligence (AI) in Professional Practice (adopted by PEO as a guideline in Ontario), Professional Engineers Ontario / Engineers and Geoscientists BC, 2024 (adopted by PEO 2025)
National Position Statement: Artificial Intelligence, Machine Learning, and Data Science, Engineers Canada, 2026
Artificial Intelligence and Consulting Engineering Practice (white paper), ACEC-Canada, 2025
Engineering Business Sentiment, H1 2026 key findings, ACEC-Canada, 2026
The Role of Artificial Intelligence in the Engineering Industry, ACEC Research Institute (American Council of Engineering Companies), 2025
Analysis on artificial intelligence use by businesses in Canada, second quarter of 2026, Statistics Canada, 2026
The Latest Deltek Clarity Industry Studies Highlight AI Challenges (47th Clarity A&E Study), Deltek, 2026
Civil engineers (21300): Canadian Occupational Projection System 2026-2035, Employment and Social Development Canada, 2026
2025 Buildings Show: Inaugural Engineering Roundtable talks technology, AI and fraud, Daily Commercial News (ConstructConnect), 2025
Secure and Governed Data Foundation for Microsoft Copilot: Foundational Deployment Guidance, Microsoft Learn, 2026
Principles for responsible, trustworthy and privacy-protective generative AI technologies, Office of the Privacy Commissioner of Canada, 2023
Bill 194, Strengthening Cyber Security and Building Trust in the Public Sector Act, 2024, Legislative Assembly of Ontario, 2024

Figures are quoted from the sources above as published. Use cases describe builds ForwardPath has delivered; client names appear only with permission.