This interview is with Lilach Bullock, AI Implementation Consultant and Fractional CMO, Lilach Bullock.
For Connectively readers, how do you introduce your role as an AI Implementation Consultant and Fractional CMO to highlight the impact you create for marketing and advertising clients?
I help businesses turn AI from a talking point into a working part of marketing.
I sit between strategy and implementation, so I question the offer, the channel, and the handoffs, then build the simplest system that addresses repetitive tasks.
After 21 years in marketing, I have no interest in adding a clever tool that creates more administrative work for the team.
The impact is a process people can use on Monday, with a human fallback when the system gets something wrong.
What key moments or choices shaped your path from traditional marketing to building AI-driven, working systems for clients?
The turning point came when I watched businesses buy more software while the work itself remained stuck. I had tested more than 300 marketing tools, and the useful ones were rarely the ones making the biggest promises. At 53, I rebuilt my own business around AI implementation and made a fairly blunt choice: I would stop handing clients plans that ended as documents. If a strategy cannot become a working system with an owner, a check, and a fallback, it is not finished.
Having tested more than 300 marketing tools, what single criterion do you use first to shortlist AI solutions for a client’s strategy?
My first question is whether the tool eliminates a repetitive task without forcing the business to reorganise itself around the software.
I have tested more than 300 marketing tools, and I trust the ones I stop noticing because they remove administrative work without demanding attention.
If a tool needs five handoffs to solve one simple problem, it is already losing.
I shortlist against the existing process first, then evaluate:
- Accuracy
- Data access
- Cost
- What happens when it fails
When you turn a repetitive marketing task into a simple workflow with a manual fallback, what is your go-to method for mapping decisions so non-technical teams adopt it quickly?
I map four things on one page:
- what starts the workflow
- what must be checked
- who makes the exception call
- what happens if the automation fails
I use the team’s own language because a beautiful diagram nobody recognises is useless.
The first test covers one job someone is already fed up with, one measure of success, and one review date.
Keeping the old manual route available for a while makes adoption much easier because nobody feels trapped by the pilot.
With a newsletter audience around 15,000, how do you translate real-time signals from replies and search terms into decisions about which segment or market to pursue next?
I pay more attention to the awkward replies than the tidy compliments. With an audience of around 15,000, a repeated question or objection is useful evidence, especially when the same wording also appears in search terms. I group those signals by the job the person is trying to do, then choose the segment with the clearest unresolved problem and a credible route to reach it. One loud reply is interesting; the same confusion appearing in the inbox and the search data is something I will build around.
For brands seeking growth without bigger ad budgets, what one content play has consistently moved the needle for you?
The content play I trust is a useful email built around a single problem people have already raised, followed by one short question.
I rebuilt an almost-dead newsletter from about an 11% open rate to the low 70s by cutting the list to people who still wanted it and refusing to write subject lines the email could not deliver on. The replies then became the next email, page, or offer. That loop has done more for growth than publishing extra content just to keep a calendar full.
In your fractional CMO engagements, what governance habit or cadence keeps AI pilots compliant, ethical, and measurable without stalling speed?
I use a short weekly review with one owner, one measure, and a visible exception log.
Before an AI pilot starts, we agree what data it can touch, which outputs need human approval, and the point at which the job is sent back to the manual process.
The review is deliberately practical; we ask:
- What did it produce?
- What was wrong?
- Who caught it?
- Did it save time or improve the result we chose at the start?
That keeps compliance and ethics inside the work instead of turning them into a presentation at the end.
What have you found most effective for earning cross-functional buy-in from sales, product, and customer success when rolling out an AI-enabled go-to-market?
I start with the handoff that all three teams already complain about.
Each team brings a different perspective:
- Sales brings the objection it keeps hearing.
- Product explains what the offer can truthfully promise.
- Customer success brings the confusion that appears after the sale.
We put one of those problems into a small pilot and show the before-and-after in their own numbers or workload.
People support an AI rollout much faster when it removes a nuisance they recognise and does not ask them to pretend the whole go-to-market process needs replacing.
Given your passion for education and social impact, how do you weave a brand’s values into AI-powered messaging so growth never compromises authenticity?
A brand’s values must show up in what the system is not allowed to say, not in a paragraph added at the end.
I lock the verified claims, proof sources, and sensitive topics before writing prompts, then maintain a human check on anything public.
If AI produces a smoother sentence that is less true, the sentence is removed. I would rather leave one real customer phrase slightly awkward than polish it into something the customer never said, because that is usually where authenticity disappears.
Thanks for sharing your knowledge and expertise. Is there anything else you'd like to add?
The useful AI conversation is no longer about collecting more tools; it’s about choosing one worthwhile job, building the system around the people who use it, and being honest about what still needs a human. That is the work I do with business owners and marketing teams. More about my AI implementation and fractional CMO work is available at https://www.lilachbullock.com.