Let me put the 3.5-day response time in context.
Industry research consistently shows that leads contacted within 5 minutes are 21 times more likely to convert than leads contacted after 30 minutes. After an hour, the probability of conversion drops by 10x. After 24 hours, that lead is essentially cold.
This firm was responding after 3.5 days. On average.
Their lead-to-meeting conversion rate was 8%. For a financial advisory firm where each new client relationship is worth $15K-50K in annual fees, every missed conversion was painful.
They had the leads. They had the expertise. They had the reputation. They just couldn't get back to people fast enough.
And the kicker? The advisors knew it was a problem. They just "didn't have time" to sit in the inbox. They were busy doing advisor work — client meetings, portfolio reviews, compliance paperwork. Lead response kept falling to the bottom of the priority list.
So the founder wrote a job description for a BDR. $65K salary plus super. The plan was simple: hire someone to watch the inbox and respond to leads.
I asked him: "What exactly would this person do all day?"
He listed:
Here's the system, step by step. No buzzwords. Just what it does.
Step 1: Instant Lead Capture When a prospect fills in the website form, HubSpot captures the data and triggers a workflow immediately. No shared inbox. No waiting for someone to notice.
Step 2: AI-Powered Response Drafting The workflow sends the enquiry details to an AI system that drafts a personalised response. Not a template. Not a "Dear [First Name]" mail merge. An actual personalised message based on:
The AI drafts the response in the firm's voice and tone. It references the specific thing the prospect asked about. It feels like a human wrote it — because a human trained the AI to write it.
Step 3: Human Review (The 2-Click Approval) The draft doesn't send automatically. It goes to the assigned advisor with two buttons: "Send" or "Edit." Most of the time, they click Send. Occasionally, they tweak a sentence. The entire review takes under 30 seconds.
This is the AI Handshake in action. AI does the heavy lifting. The human provides the final judgement.
Step 4: Automatic Lead Routing Based on the enquiry type and the advisor's specialisation, the lead is automatically assigned to the right person. No manual triage. No "who should handle this one?" meetings.
Step 5: Follow-Up Sequence If the prospect doesn't reply within 48 hours, an automated follow-up sequence kicks in. Three touchpoints over two weeks, each slightly different, each adding value rather than just "checking in."
Step 6: CRM Updates Every interaction — the initial response, the follow-up, the meeting booking — is automatically logged in HubSpot. No manual data entry. The advisor's CRM record is always current.
Step 7: Escalation If a lead goes cold after the sequence, a task is created for the advisor to make a personal phone call. This is where human judgement takes over. AI handled the first 90% of the process. The human handles the nuanced last 10%.
The average response time went from 3.5 days to 18 minutes. Not because the advisors suddenly became faster — because the AI drafted the response instantly, and the advisor just had to approve it.
Most approvals happened within 15-20 minutes of the enquiry. During business hours, some happened in under 5 minutes.
The conversion rate nearly tripled. The single biggest factor? Speed. Prospects were genuinely surprised to get a personalised, relevant response within minutes of enquiring. Several commented: "I can't believe how fast you got back to me."
Before the system, CRM data was patchy. Advisors would forget to log emails, skip updating deal stages, and rarely add notes. With the automated system, every interaction was logged automatically. The founder could finally see an accurate pipeline for the first time.
Each advisor was spending roughly 45 minutes per day on lead response and follow-up. That dropped to about 10 minutes (reviewing AI-drafted responses). That's 35 minutes per advisor per day — or about 2.5 hours per day across the team. Over a year, that's 600+ hours redirected to actual client work.
The AI system costs are minimal: workflow automation is included in their HubSpot subscription. The AI response drafting costs about $100-150/month in API calls. Total annual cost: under $2,000.
Versus the $65K+ hire they were planning.
Note: There was also a $5,000 implementation cost, and a $3,000 hardware cost. Still cheaper then $65K.
And here's where I earn your trust. Because if I only told you the wins, I'd be no better than every "AI is magic" LinkedIn post you've already scrolled past.
A prospect sent an enquiry that included: "My husband just passed away and I need to understand our financial situation."
The AI drafted a response that was technically perfect — mentioned their financial planning services, offered a meeting, was professional and warm. But it didn't pause. It didn't lead with empathy. It jumped to logistics too quickly.
A human would have read that message and written something very different. Something that acknowledged the grief before mentioning services.
We caught it during the review step (the advisor edited before sending). But it was a wake-up call: AI can mimic tone. It can't feel context.
The fix: We added flag keywords — "passed away," "divorce," "illness," "emergency" — that trigger a different workflow. Instead of drafting a response, the system alerts the advisor directly: "Sensitive enquiry. Please respond personally." No AI involvement.
Some enquiries don't fit into neat categories. A prospect asking about "restructuring my SMSF to include my new business partner while also setting up insurance for my kids" — that's not something an AI can confidently route or respond to.
The AI would draft something generic. The advisor would have to rewrite it entirely, which took longer than writing from scratch.
The fix: We built a complexity scoring system. If an enquiry contains multiple topics or unusual keywords, it bypasses the AI entirely and goes straight to the advisor as a "manual response needed" task.
Our initial automated follow-up sequence sent three emails in seven days. For financial services — where trust is everything — that felt pushy. Two prospects specifically mentioned feeling "hounded."
The fix: We stretched the sequence to three touchpoints over three weeks. And we made each follow-up add genuine value: a relevant article, a free resource, a case study. Not "just checking in" — actual reasons to engage.
After three months, the advisors had gotten used to the AI handling lead response. When the system had a brief technical issue (an API timeout), nobody noticed for four hours. Leads went unresponded.
The fix: We built a monitoring alert: if any lead goes unresponded for more than one hour during business hours, the founder gets a direct notification. Belt and braces.
Let me be straight with you — the “$65K savings” headline is only half the story.
On paper, they were about to commit to a $65K salary, plus super and on-costs that pushed the real annual outlay to roughly $78K. And that’s before you factor in recruitment fees, the time cost of interviews, onboarding, and the inevitable dip in productivity while a new hire finds their feet.
By contrast, the AI-led system came in at around $2K a year in running costs — mostly API usage and automation that was already covered in their existing HubSpot subscription. No sick leave. No annual leave. No performance management. Just a predictable, low operating cost that scales with usage instead of headcount.
There were some upfront investments: about $5,000 in implementation and another $3,000 in hardware (Local AI Model). That covered scoping the workflow, building the automation, integrating the AI model, testing edge cases, and setting up a reliable review and monitoring layer so advisors stayed in control. Even when you add those in, you’re still well under the ongoing $65K salary line in Year 1, and the gap only widens in Year 2 and beyond as the one-off costs fall away and the system keeps performing.
And then there’s the hidden upside a salary line never shows you: the AI doesn’t take 3–6 months to ramp up. It was delivering 18-minute response times from day one. It doesn’t cap out at a certain number of leads per day. It doesn’t resign just as it becomes brilliant at the job. It quietly does the 70% of work that’s repeatable, every single day, so the humans can focus on the 30% that actually moves the needle.
After building this system — and after running Kit as our own AI employee at DigiKat — I've developed a clear framework for where AI should end and humans should begin.
I call it the AI Handshake.
AI excels at:For the financial advisory firm, the handshake happens at the response drafting stage: AI drafts, human approves. For DigiKat, the handshake happens later: Kit handles entire workflows autonomously, and humans step in only for strategic decisions.
Your handshake point depends on:
Automate the repeatable. Personalise the valuable.
Without hesitation. Yes.
The financial advisory firm saved $78K in hiring costs, generated $140K in additional revenue, and got their advisors 600+ hours back. The system isn't perfect — we've patched the emotional nuance issue, the complexity routing, and the follow-up cadence. There will be more patches.
But the question isn't "is AI perfect?" The question is "is AI better than the alternative?"
The alternative was a $65K hire who would still take 3-6 months to ramp up, still need leave and sick days, still max out at handling a fixed number of leads per day, and still forget to update the CRM sometimes.
The AI system was responding in 18 minutes from day one. It hasn't taken a sick day. It handles as many leads as come in. And it never forgets to update the CRM.
Perfect? No. Better than the alternative for 70% of the work? Absolutely.
The other 30%? That's what the humans are for. And they're better at that 30% now because they're not buried under the 70% that AI handles.
That's the AI Handshake. That's the honest truth.
If you're considering hiring for a role that involves monitoring, responding, routing, data entry, or reporting — pause.
Ask yourself: what percentage of this role is repeatable? If it's over 50%, there's probably an AI-assisted workflow that handles the repeatable part, freeing a human to handle the valuable part.
You might not save $65K. You might save $30K. Or you might just give your existing team 10 hours a week back. Either way, the ROI is there.
But do it honestly. Know where the AI Handshake is for YOUR business. Don't automate the human stuff. Don't leave the robot stuff to humans.
Automate the repeatable. Personalise the valuable.If you want to explore what an AI-assisted workflow could look like for your specific situation, reach out. No pitch. Just an honest conversation about what's automatable and what's not.
Because the honest truth? AI isn't replacing humans. It's replacing the boring parts of human jobs. And that's a trade worth making.
Mick Goman is the founder of DigiKat. His AI employee Kit runs daily operations. His human employees run the relationships, strategy, and creative work that AI can't touch.