The SERP row nobody checks
Consulting client follow-ups that quote your last call, in your voice.
Every tool on the first page of Google for this keyword is a CRM with a drip sequence and merge fields. None of them read the transcript of the call you just finished. None of them compose the draft in the phrasing Clone has observed in your outbox. The Memory layer does both, then hands you a draft in Gmail that sounds like you wrote it on a good day.
The first page of Google, summarized
Twelve cadence tools. All merge fields. Zero transcript reading.
Every top SERP result is either a CRM with a sequencer bolted on or a dedicated cadence tool. Great at pausing on reply. Blind to what the client actually said on the Tuesday call.
The core claim
A good follow-up has context and voice. Merge-field drips have neither.
Think about the last follow-up you actually liked receiving. It quoted something specific you said. It was written in the sender's normal phrasing, not in sequencer-English. It showed you the sender was still in the room, mentally, two days after the call.
Now compare that to what the CRMs on the first page of Google produce. Hi {{first_name}}, great connecting on {{last_meeting_date}}, wanted to follow up on the opportunity at {{company}}. Your client reads that and immediately knows they are in a cadence. The follow-up becomes a tax, not a touch.
Clone takes a different path. Two inputs: the transcript of the call (from whatever tool you already record with) and your last ~12 sent emails at the matching stage. One output: a Gmail draft that quotes the call and sounds like your outbox. No merge fields. No generic body. Nothing a CRM sequence could produce.
Side by side
The same follow-up moment. Two very different emails.
You ran a 45-minute Zoom with Priya at Holloway on Tuesday afternoon. Three commitments surfaced. She needs a recap on Wednesday. Here is what the SERP answer drafts, and here is what Clone drafts.
Hi {{first_name}}, Great connecting with you on {{last_meeting_date}}. As discussed, I wanted to follow up on the opportunity at {{company}}. Let me know if you have any questions. Best, {{sender_first_name}}
- Merge fields visible in the body (or visible when one fails)
- No reference to anything actually said on the call
- Tone is generic SDR, not your tone
- No conditional rule, no escalation, no cc logic
Where the context and voice come from
Four sources in. One draft out. None of them are merge fields.
Inputs → Clone Memory → the draft that lands in Gmail
The ten-second loop after every call
From 3:00pm Zoom to 3:52pm draft. No middleware.
The sequence is the same for every follow-up stage. The transcript tool changes, the voice source changes, the draft stays a draft until you approve it.
post-call follow-up sequence
Watching Clone run it, once
The first run after your next call, line by line.
This is what the chat log looks like on the first real draft after a call with an existing client. Nothing is sent. Everything is logged. The transcript and voice sources are named so you can see exactly where each phrase came from.
The anchor: memory/rituals/followup.md
One markdown file. Five stages. No dashboard.
This is the file that defines which follow-up stage pulls from which source, which voice sample to mimic, and which conditional rules apply. If you want a stage to pull from Fireflies instead of tl;dv, or to skip if the client replied in the last 30 days, you change one line here. There is no UI. That is the point.
By the numbers
The whole mechanism in four numbers.
On the voice side
The 12-email window comes from the pattern described in src/components/how-it-works.tsx step 03: Clone observes a pattern across your last dozen same-stage emails, surfaces it for confirmation, saves it to Memory, reuses it. Twelve is enough to catch the openers and sign-offs you use every time. Fewer and the pattern is noisy.
On the context side
The four transcript tools (tl;dv, Fireflies, Otter, native Zoom) all work the same way for Clone: it reads the transcript page on screen. You pick whichever you already use. Switching is a one-word edit of the ritual file. The transcript never flows through a Clone server.
The uncopyable detail
A conditional rule Clone observed, not a path you built.
The specific example in the product's own how-it-works panel (src/components/how-it-works.tsx lines 44-56) is worth quoting directly, because it is the rule most CRM sequence tools cannot express without making you build a three-node path in Zapier:
Noticed a pattern in your last 12 kickoff emails:
• You start with a personal line
• You always attach the SOW PDF
• You cc your assistant when the contract is above $10K
Should I apply this template going forward?
That is the mechanism. Clone watches your outbox, surfaces a pattern, asks once, saves to Memory. The rule then applies every time the stage triggers, forever, until you edit the markdown file. In HubSpot Sequences, you would have to build that cc-rule as a separate workflow with a value-comparison branch. In Outreach or Salesloft, it would be a template tag. In Clone, it is one sentence you confirmed once, against a rule Clone watched you follow in your own sent mail.
This is the anchor of the whole page. Every other thing Clone does for follow-ups inherits from this one behavior: observe patterns in your actual past work, confirm once, apply forever, and let you edit the markdown file if the rule changes.
“I tried HubSpot Sequences for six months before I realized my clients were starting every reply with 'I know this is automated but'. Clone is the first tool where nobody has said that. The follow-ups just sound like me, because they are stitched together from the calls and the emails I was going to write anyway.”
Setup in four steps, first draft in under fifteen minutes
From signup to a real follow-up draft after your next call.
Point Clone at your transcript source
tl;dv, Fireflies, Otter, Granola, or your native Zoom cloud recordings. One of those is already running on your calls. Clone reads from whichever you pick. No plugin swap required.
Let Clone read your sent-mail folder once
Clone observes patterns across your last ~12 emails per follow-up stage. Your opener, your sign-off, whether you bullet commitments back, whether you use em dashes or commas, whether you cc an assistant when the contract is above a threshold.
Write memory/rituals/followup.md
Five to seven stages that match how your practice actually works. Discovery → proposal → mid-engagement pulse → post-engagement → reactivation → ghosted. Eight to twenty lines total. Plain markdown. Any text editor.
Run the first real follow-up while watching
After your next call, Clone opens Gmail, drafts the follow-up using the transcript and your observed voice, and posts the draft link in the chat. You read it. You fix the one line that was off. That fix goes back into Memory.
When this guide applies to you
If three or more of these are true, keep reading.
Signs the merge-field approach is costing you replies
- You have tried Copper / Pipedrive / HubSpot sequences and your clients can smell them
- You record every call on tl;dv or Fireflies but nothing in your CRM touches the transcript
- Your follow-ups sound different depending on whether it is Monday or Friday night
- You cc an assistant on contracts above a threshold and keep forgetting
- You have a dozen ghosted proposals you never nudged because the template felt wrong
- You want the follow-up to quote the call, not reference 'our last conversation'
- You want client data to stay on your Mac, not flow into another vendor cloud
- You are not ready to rip out your CRM to get this, you just want the drafts to land
Clone vs the SERP, one row at a time
Feature comparison against the top SERP answers
Every row is a capability either present or absent. None of the comparison is about preference, it is about what the tool will actually produce on Wednesday morning after a Tuesday call.
| Feature | CRM sequencers (Copper, Pipedrive, HubSpot, Outreach, Salesloft) | Clone |
|---|---|---|
| Pulls context from your last Zoom call transcript | Out of scope. CRM sequences pull from the CRM record (stage, last-touch date, deal value). The call transcript, if you record one at all, lives in a separate tool and never gets composed into the follow-up body. | Clone reads tl;dv, Fireflies, Otter, or native Zoom cloud recordings and lifts 2-3 specific commitments from the transcript. Those commitments go into the draft verbatim, as quotes. |
| Mimics your actual writing voice | Merge fields. {{first_name}}, {{company}}, {{last_meeting_date}}. The body is a template written once by whoever set up the sequence. It sounds like every other SDR on the planet. | Clone observes patterns across your last ~12 sent emails per follow-up stage. Your opener, your bullet style, your sign-off, your em-dash-vs-comma habit. The draft sounds like you wrote it, because it was written from your sent mail. |
| Applies conditional rules you already follow | Condition branches exist in Zapier and HubSpot, but they require you to build them explicitly: trigger, filter, path, action. You have to know every rule in advance. | Clone infers rules from behavior. 'You cc your assistant when the contract is above $10K' is observed, surfaced once for confirmation, saved to Memory. No path builder. No trigger. You say yes once. |
| Works without replacing your CRM | Copper, Pipedrive, HubSpot, Outreach, Salesloft all require you to be in their CRM. Switching CRMs means rebuilding every sequence, every template, every merge field. | Clone drives Gmail, your current CRM (HubSpot, Pipedrive, Folk, or a Google Sheet), and whichever transcript tool you use. Switch CRMs and edit one line of rituals/followup.md. Nothing to rebuild. |
| Works without an API key for the transcript source | Zapier and iPaaS tools need an API integration for tl;dv / Fireflies / Otter. Paid tiers only on most. | Clone's Computer Agent layer reads the transcript page on screen. No API. No OAuth. Works on the free tier of every transcript tool. |
| Pauses when the client replies, resumes if they go quiet again | Most drip tools pause on reply. Very few resume correctly if the client goes quiet a second time, and none of them handle 'they replied but did not answer the ask' gracefully. | Clone reads the reply, checks whether the ask was answered, and either stops the sequence or drafts a short clarification in your voice if the reply dodged the question. |
| Monthly cost for a solo consultant | Copper $29-134/seat. Pipedrive $14-99/seat. HubSpot Sales Hub $20-1200/mo. Outreach $100+/seat. Salesloft quoted. Plus the cost of the sequences-add-on tier in each. | $49 per month on Solo. One plan. The transcript tool you use is whatever you already pay (tl;dv free tier is enough). Switching transcript tools is a one-word edit. |
Common questions about consulting client follow-up automation
How does Clone actually lift a 'commitment' from a call transcript?
The Computer Agent layer opens the transcript page for the call (tl;dv, Fireflies, Otter, Granola, or your native Zoom cloud transcript) and reads the content on screen. The Planner layer classifies sentences in the transcript into commitments, decisions, risks, and small talk. A commitment is a sentence where one party said they will do something by a date, or asked the other party to. Those are the lines Clone quotes in the follow-up draft. The rest of the transcript stays in the transcript tool. You can see exactly which lines Clone lifted because the draft review panel shows the timestamp ranges it pulled from.
What does 'learns your voice from your last 12 emails' actually mean?
It is the same mechanism described in the how-it-works section of the site (src/components/how-it-works.tsx lines 43-58). Clone reads your sent-mail folder and observes patterns across your last dozen replies at each follow-up stage. That includes the structural patterns (a single personal-line opener, a bullet list of commitments, a specific sign-off) and the conditional ones (cc your assistant when the contract is above $10K). The first time Clone sees a clear pattern, it surfaces it in the chat: 'Noticed a pattern in your last 12 proposal follow-ups. Should I apply it going forward?' You say yes once. The pattern is saved to the Memory layer (src/components/architecture.tsx line 25) and reused. No prompt-tuning, no style guide, no brand voice configuration. It is your own outbox, observed.
What is in memory/rituals/followup.md and why does it matter?
It is the plain-markdown file that maps each follow-up stage to the source Clone pulls context from, the voice source it should mimic, and any conditional rule (like 'cc assistant if contract > $10K' or 'skip if replied in last 30 days'). Five to seven stages for a typical consulting practice: discovery call, proposal sent, mid-engagement pulse, post-engagement reactivation, ghosted proposal nudge. Eight to twenty lines of markdown. The file is the source of truth. Editing it is how you change the behavior. There is no dashboard, no path builder, no sequence editor. If you want a stage to start pulling from Fireflies instead of tl;dv, you change one word in the file.
How is this different from HubSpot Sequences, Outreach, or Salesloft?
HubSpot / Outreach / Salesloft are cadence tools built around a CRM stage and a schedule. The content of the follow-up is a template with merge fields ({{first_name}}, {{company}}, {{last_meeting_date}}) written once by whoever set up the sequence, then stamped out. They pause on reply and resume on silence, which is good. But the body of the email has no awareness of what was actually said on the call, and the tone is whatever the template author wrote, not yours. Clone takes the opposite approach: the cadence rules stay simple (send after call, nudge at day 7, etc.), and the intelligence goes into the content, pulled from the transcript and the voice of your sent mail.
How is this different from the Clone milestone follow-up automation guide on this site?
That guide (see /t/consulting-milestone-followup-automation) is about the cadence of follow-ups after a deliverable: day 3, day 7, day 14 with timeline-shift language. This guide is about the content of every follow-up moment across the client lifecycle: discovery, proposal, mid-engagement, post-engagement, reactivation, ghosted nudge. The two are orthogonal. A mature consulting practice runs both at once: the day 3/7/14 cadence from the milestone guide, with each message composed using the transcript-plus-voice mechanism described here. Both rituals live side-by-side in memory/rituals/.
Does Clone actually work with the free tier of tl;dv, Fireflies, and Otter?
Yes, because the Computer Agent layer reads the transcript page on screen rather than going through an API. There is no paid-tier integration gate. If you can open the transcript in a browser tab, Clone can read it. The architecture principle is in src/components/architecture.tsx lines 56-59: 'Clone uses the apps you already pay for. Switch CRMs, change invoicing tools, add a new client portal, Clone adapts in the same conversation. No re-wiring required.' That applies to transcript tools as well. Most solo consultants use tl;dv free (unlimited recordings and AI summaries), and Clone works against that tier without any upgrade.
What does the first week of using this actually look like?
Day one: install the app, write eight to twelve lines of memory/rituals/followup.md with your stages, run the ritual manually once after your next call. You watch Clone open the transcript, read it, draft the follow-up in your voice, and post a review link in the chat. You fix the one line that is off and the fix goes back into Memory. Day two to five: every post-call follow-up gets drafted automatically within five minutes of the call ending. You review, tweak, and send. By end of week one, Clone has observed enough of your edits that drafts start landing right on the first pass for your most common stage. That is when you consider flipping auto_send on for that one stage. Contracts above your review_first threshold (default $10K) stay manual review always.
What about GDPR, data residency, and client confidentiality?
Client data never leaves your Mac. The transcript stays in your transcript tool. The email draft stays in Gmail. The rituals file stays on your disk. This is the first architectural principle (src/components/architecture.tsx lines 46-50): 'Clone operates your desktop apps from your desktop. Client files, emails, contracts, and transcripts never leave your computer. Your engagements stay confidential by default.' The Planner layer may call a model to interpret your English instruction, but the attached client data, the transcript, the sent-mail corpus are not sent along with it. If your client says 'no AI on our calls', you simply do not record. If your client says 'no cloud transcripts', you use the native local recording option in Otter or Granola.
What happens when Clone drafts something that is wrong?
Every draft lands in Gmail as a draft, not as a sent message, on the first pass and on review_first stages. You see the draft. You fix the line that was wrong. The fix goes back into Memory as a voice-pattern update, so the next draft at that stage learns from your correction. The 'Always reviewable' principle (src/components/architecture.tsx lines 61-63) applies literally here: every action Clone takes is logged and reversible, and you can roll back an entire morning of drafting with one click if you need to. For most consultants, the single most common correction in week one is a phrasing edit ('you said it warmer than that') and by week three those corrections disappear.
Can I run this without any CRM at all?
Yes. A lot of solo consultants run a single Google Sheet as their CRM. Clone works against a Sheet the same way it works against HubSpot or Pipedrive: it reads the row, writes the next-step field, updates the stage. The ritual file points at whichever CRM you actually use. If you have no CRM, you declare the Sheet as your CRM in the file, and Clone treats it as one. The point of the rituals/followup.md file is that the CRM is a source of truth for the stage, not a required piece of software. See the tool-agnostic principle (src/components/architecture.tsx lines 56-59). The follow-up still pulls its content from the call transcript and your sent mail, either way.
How much does it cost?
$49 per month on Solo, one plan, 21-day free trial. No per-email fee. No merge-field tier. No integration-seat tier. Compare against Copper $29-134/seat, Pipedrive $14-99/seat (sequences are on higher tiers), HubSpot Sales Hub $20-1200/mo, Outreach $100+/seat (quoted), Salesloft quoted. None of those include the transcript tool. tl;dv free tier is enough for most solos, and Clone reads it without an API key.
Other guides about the operational layer of a consulting practice
Keep reading
Consulting Milestone Follow-Up Automation
The cadence side of the story: day 3, day 7, day 14 with timeline-shift language, decoupled from your client's inbox.
Invoicing Automation Software Without an API
The same Computer Agent layer, pointed at QuickBooks Desktop, a Google Sheet, or your practice-management billing tab.
AI Assistant for Small Business
Why a plain-English AI that drives your existing apps beats a SaaS replacement stack for small teams.
Try it on one stage this week
Pick your most common follow-up moment. Let Clone draft it once.
Install the app. Write four lines of memory/rituals/followup.md for that one stage. Run it once after your next call. Read the draft. If it sounds like you, you already know. $49 a month on Solo, 21-day free trial.
$0/mo on Solo · no merge fields ever · data stays on your Mac