Pearl Lemon Group’s AI reality: automation with human review
Deepak Shukla, founder of Pearl Lemon Group, describes a business spanning marketing, lead generation, specialist services and in-house software. This conversation looks at where AI can handle first-pass work, why unusual client decisions still need human review, and how the group separates useful customer interest from traffic.
Use AI for first-pass triage, research and anomaly spotting, but keep a person accountable for exceptions such as unusual briefs, delivery promises and pricing. Measure qualified enquiries and downstream outcomes rather than impressions or AI usage.
What’s the quick origin story of your business, and what makes what you do genuinely different from the other options your customers are looking at?
I started with SEO and customer acquisition. Over time, running an agency made it obvious that generating interest and successfully delivering the work were different disciplines, so Pearl Lemon expanded into specialist service businesses and our own software tools. The practical difference is that we have to use many of these systems inside our own operating businesses; a dashboard can look impressive while a real customer still goes unanswered.
Where does most of your new business come from today, and where do you wish it came from?
Organic search, direct enquiries and introductions have long been important ways people find our businesses, though the mix varies by service. I’d like more enquiries from people with a specific problem, realistic budget and clear timetable—not simply a bigger volume of loosely qualified contacts. We look at whether interest turns into a serious conversation, not just whether impressions are up.
What’s one thing you’ve actually handed over to AI in the last year, and one thing you tried it on and went back to doing the human way?
We’ve handed AI some of the first-pass work: assessing what an incoming enquiry is about, drafting research, and surfacing SEO or operational anomalies for review. I’ve also tried treating a rapid automated answer as if it were a finished answer, and gone back to a human review whenever an unusual client brief, delivery promise or pricing judgment is involved. Automation is most useful when somebody is accountable for correcting exceptions.
Have you noticed a change in how customers find you or what they already know before they reach out? Is anyone arriving through AI assistants like ChatGPT yet?
Buyers increasingly arrive with quite specific questions, but I wouldn’t claim every well-informed enquiry came from ChatGPT. Across our Search Console portfolio, the settled week of 12–18 October showed roughly 7% more search impressions than the previous week while clicks fell roughly 9%; that tells me attention and useful action are not the same thing. We need proper referral attribution before claiming AI assistants generated a particular volume of leads.
If you were advising someone in your industry on all of this for the next 12 months, what would you tell them to actually do, and what would you tell them to ignore?
Start by measuring the business outcome you want: qualified enquiry, booked meeting, sale or retained customer. Make existing pages answer the most specific commercial questions your audience actually searches for; fix broken handovers and mobile forms before producing another hundred AI-written pages. Use AI to organise information and spot anomalies, but audit what actually shipped. Ignore vanity usage statistics and promises that an autonomous agent can replace judgment in every customer interaction.
Thank you to Deepak Shukla and the team at Pearl Lemon Group for sharing what actually worked with Leaders Perception readers.
Want to share your playbook?
Leaders Perception publishes founder and operator interviews every week. A few questions, about five minutes, no calls. It is free and everyone we accept gets published.