Marketing Automation in 2023: AI Drafting and CRM Sync

By 2023, marketing automation had matured past being a single tool bolted onto an email platform. It had become the connective layer between marketing, sales and the data both teams relied on, spanning email, SMS, ad retargeting and the CRM. This playbook covers how to run that layer well, not just turn it on.

Start with the customer journey, not the software

The most common mistake teams make is opening an automation platform and building flows around what the tool makes easy, rather than mapping the actual path a customer takes first. Before building anything, sketch the real stages: how someone finds you, what makes them consider buying, what happens right after they buy and what brings them back. Automation should follow that map, not replace the thinking behind it.

Build lead scoring that reflects real intent

A lead who downloaded a single guide and never opened another email is not the same as one who’s visited your pricing page three times this week. Score leads on a combination of engagement (opens, clicks, page visits) and fit (company size, role, stated need if you collect it), and route only leads crossing a real threshold to sales. Automation that hands every form-fill straight to a salesperson wastes their time and trains them to ignore what marketing sends over.

Connect the CRM properly, not just superficially

Automation that lives only inside an email or marketing tool, disconnected from what your sales team sees, creates two versions of the truth. Set up genuine two-way sync: automation should read stage and ownership data from the CRM, and actions taken in automation (opened a demo request, clicked a pricing link) should write back to it. This is the difference between automation that supports sales and automation that quietly works against it.

Use AI drafting tools as a first draft, not a final one

Following the broader adoption of generative AI tools through 2023, most marketing automation platforms added AI-assisted content generation directly into their workflow builders, drafting subject lines, email body copy and even basic ad variants. Used well, this cut the time to build a new flow significantly. Used carelessly, it produced generic copy that sounded like every other brand using the same underlying model. The teams that got value from it treated AI output as a fast first draft that a human still edited for voice, accuracy and specificity before anything shipped.

Build cross-channel journeys, not parallel ones

Email and SMS running as two separate, uncoordinated systems create a real risk of over-messaging or contradicting yourself across channels. A well-built journey treats channel as one variable in a single flow: a cart abandonment sequence that starts with email and escalates to SMS only if the email goes unopened, for instance, rather than firing both independently on their own schedules.

Report on outcomes your finance team would recognize

Automation platforms report dozens of vanity metrics by default. Build your actual reporting around revenue per flow, cost per acquired customer through automated channels and retention lift from lifecycle sequences. These are the numbers that justify continued investment in the platform and headcount to maintain it, and the ones that let you make an honest case for cutting a flow that isn’t earning its keep.

Maintain what you build

An automation flow built once and left alone degrades as products change, pricing shifts and the audience it was designed for moves on. Put a recurring review on the calendar, quarterly at minimum, to check performance against when the flow was first launched and update copy, offers and triggers accordingly. Automation isn’t a project with an end date; treating it as ongoing infrastructure is what separates the businesses still getting value from it years later from the ones sitting on a pile of stale, forgotten workflows.