Automation
Marketing automation: what to automate first, and what to keep human.
Automation should give your team time back for the work only people can do. Done in the wrong order, it just sends the wrong message to more people, faster. Here's how to sequence it.
Most marketing teams don't have an automation problem. They have a sequencing problem. Tools get bought for the exciting use case, like personalized journeys or AI-written campaigns, before the basics underneath them work. Then the data is messy, the reports don't match and nobody trusts the output.
A better marketing automation strategy starts with a simple question: what work is repetitive, predictable and low risk if a machine does it? Automate that first. Protect the work that depends on judgement.
What the data says about automation today
Automation and AI are now standard. HubSpot's 2026 State of Marketing report, based on more than 1,500 marketers, found 86.4% of marketing teams use AI in at least a few areas. Administrative task automation was among the top uses, with 35.6% using AI extensively for it and another 40.5% occasionally. About a third of respondents said AI saves their team 10 to 14 hours a week.
The gap is in execution. Salesforce's 2026 State of Marketing research, a survey of 4,450 marketing decision-makers, found 75% of marketers have adopted AI, yet 84% admit to running generic campaigns and 69% struggle to respond to customers promptly. High performers were 2.4 times more likely to have unified their data sources.
Much of the automation teams pay for also goes unused. Gartner's 2025 Marketing Technology Survey found marketers use 49% of their martech stack's capabilities, up from a low of 33% in 2023 but below 58% in 2020 (as reported by CMSWire, 2026). And in MarTech's 2025 State of Your Stack survey, 65.7% of respondents named data integration their biggest stack management challenge, and 45% cited a lack of skilled resources.
Data quality is the thread running through all of this. HubSpot found just 65% of marketers say they have high-quality audience data, a figure that didn't move year over year. Automation built on poor data scales the problem.
A note on the data
These are vendor and industry surveys of self-selected respondents, mostly outside Canada, and MarTech's survey had 169 completed responses. Time-saving figures are self-reported. Treat them as directional rather than precise benchmarks for your business.
What to automate first
In rough priority order, based on time saved, risk and how much later work depends on it:
- Reporting. Recurring weekly and monthly reports are the easiest win. Connect sources into one dashboard, agree definitions once and schedule delivery. It also exposes tracking gaps early. Our article on why your marketing numbers don't add up explains what to check.
- CRM data hygiene. Automate deduplication, field standardization, required fields on forms and alerts for stale records. Every other automation depends on this.
- Lead routing and assignment. Send leads to the right person instantly based on clear rules such as region, product or company size, with alerts when nobody follows up. Speed and consistency matter more here than sophistication.
- Trigger-based emails. Confirmations, welcome series, event reminders and simple post-purchase follow-ups. Predictable, useful and easy to test.
- Lead nurture. Longer sequences based on behaviour and stage. Worth doing once your data is clean and you know what content actually moves people forward.
- Scoring and personalization. Valuable, but only once the first five are working. Scoring built on bad data sends sales the wrong leads with false confidence.
What to keep human
- Positioning and strategy. Who you're for, why you win and where you'll spend. Tools can surface data, but they can't make the trade-offs. This is the heart of building the plan before buying the tools.
- Creative judgement. AI can draft and vary, but deciding what's on brand, what's genuinely interesting and what might land badly takes a person.
- Final approval of public content. Anything that goes to customers or the public should have a named human who signed it off.
- Key relationships. Top customers, partners and high-value prospects deserve a real person. HubSpot found 73.4% of marketers see AI working in conjunction with marketers, not replacing them.
- Exceptions and complaints. Anything outside the rules, or anything emotional, should route to a person quickly.
| Automate | Keep human | |
|---|---|---|
| Nature of the work | Repetitive, rules-based, high volume | Judgement, trade-offs, context |
| Cost of an error | Low and easy to fix | High, public or hard to reverse |
| Examples | Reports, routing, data clean-up, confirmations | Positioning, creative, approvals, key accounts |
| Role of people | Set rules, monitor, improve | Decide, with automation preparing inputs |
CASL and automated email in Canada
Automation doesn't change your legal obligations. Canada's Anti-Spam Legislation (CASL) applies to commercial electronic messages whether a person or a workflow sends them. The government's guidance on getting consent to send email sets out the essentials:
- Consent. You need express or implied consent. Express consent doesn't expire unless withdrawn. Implied consent is generally time-limited: up to two years from a purchase, or six months from an inquiry.
- Identification. Clearly identify your business, with a mailing address and a phone number, email or website.
- Unsubscribe. Include a working unsubscribe mechanism, valid for at least 60 days after sending, and honour requests within 10 business days at no cost to the recipient.
- Records. Be ready to prove consent. This is where automation helps: capture the source, date and wording of consent in your CRM.
Build implied-consent expiry dates into your workflows so contacts drop out of sequences automatically. Under section 20 of the Act, the maximum administrative penalty is $1 million for an individual and $10 million for any other person. This is general information, not legal advice. Exemptions and details matter, so get qualified legal advice before launching automated programs.
A 30/60/90-day rollout
Days 1 to 30: Foundations
Audit what you already own and what's switched off. Clean and standardize CRM data, document consent sources and set up automated reporting. Agree on lifecycle stages with sales. Our performance marketing audit checklist covers the tracking side.
Days 31 to 60: Core workflows
Launch lead routing with follow-up alerts, transactional and welcome emails, and data hygiene rules. Test every workflow with internal contacts first. Measure response time and data quality against your day-one baseline.
Days 61 to 90: Nurture and refine
Add one or two nurture journeys tied to real buying stages. Start simple lead scoring if data allows. Review what's working, switch off what isn't, and decide what comes next. Removing unused tools is often where budget leaks get plugged.
One more data point worth weighing: MIT NANDA's 2025 report found that while roughly half of generative AI budgets went to sales and marketing, some of the clearest returns came from back-office automation. Unglamorous work is often where the payback is.
How Three Dots Brands approaches this
We help brands in Toronto and across Canada plan and roll out marketing and CRM automation using our Think, Plan, Execute approach. We start by finding where your team's time and data are leaking, decide what's worth automating and in what order, then work with your team or partners to build it. We don't resell software, so our advice isn't tied to a platform. If you'd like a second opinion on where to start, try our marketing diagnostic, or see whether a fractional CMO is a better fit for your stage.
Frequently asked questions
What should a business automate first in marketing?
Start with work that is repetitive, rules-based and low risk if it goes wrong: recurring reporting, lead routing and assignment, CRM data hygiene, and simple trigger-based emails such as confirmations and welcome series. These save time quickly and create the clean data that more advanced automation depends on.
What parts of marketing should stay human?
Positioning, strategy and budget decisions, creative judgement, final approval of anything public-facing, and relationships with key customers and partners. Automation can support these by preparing data and drafts, but the decisions and the accountability should sit with people.
Does CASL apply to automated marketing emails?
Yes. Canada's Anti-Spam Legislation applies to commercial electronic messages whether they are sent manually or by an automated workflow. You need express or implied consent, clear sender identification and contact details, and a working unsubscribe mechanism that is honoured within 10 business days. Rules have specific exceptions and time limits, so get qualified legal advice for your situation.
How long does it take to implement marketing automation?
A focused first phase covering data clean-up, reporting, lead routing and one or two core email journeys can usually be set up over about 90 days. More complex lead scoring, personalization and multi-channel journeys take longer and should be built on top of that foundation once it is working.
Sources
- HubSpot, 2026 State of Marketing: Data from 1,500+ global marketers, 2026.
- Salesforce, 75% of Marketers Have Adopted AI Yet Still Use It To Send One-Way, Generic Campaigns (State of Marketing, tenth edition), 2026.
- CMSWire, Why Are Marketers Still Paying for Martech They Never Use? (citing Gartner's 2025 Marketing Technology Survey), 2026.
- MarTech, These are the challenges and barriers impacting your martech stack (2025 State of Your Stack), 2025.
- MIT NANDA, The GenAI Divide: State of AI in Business 2025 (hosted copy of the report PDF), 2025.
- Innovation, Science and Economic Development Canada, Getting consent to send email, current guidance.
- Government of Canada, Justice Laws Website, An Act to promote the efficiency and adaptability of the Canadian economy (Canada's Anti-Spam Legislation), S.C. 2010, c. 23, current to date.
Gartner figures are cited through CMSWire's reporting of Gartner's survey. Industry surveys are self-reported and largely global rather than Canadian. CASL details are summarized from official sources and are not legal advice.