AI Agents and Assistants
What is an AI agent? A plain-English guide for business owners
What an AI agent is, how it works, where it helps a small business, the risks, and a 7-point check to see if you are ready for one.
Supervised Agents
Human-in-the-Loop Sales Agent
Drafts every sales message and action; a rep approves before anything reaches a prospect.
Your reps lose hours writing follow-ups and proposals, but you are not ready to let AI email prospects on its own. This agent qualifies leads and drafts the next message, follow-up or proposal from your sales playbook. Nothing reaches a prospect until the rep you name approves, edits or rejects it, and rejected drafts come back better next time.
- Outreach, follow-ups and proposals drafted from your sales playbook
- No message reaches a prospect without a rep's approval
- Rejection reasons fed back so later drafts improve
$5,400
20–30 working days
Optional maintenance $270/mo
What it is
Human-in-the-Loop Sales Agent qualifies leads and drafts the outreach, follow-up or proposal that comes next, but nothing goes to a prospect until the rep you name approves, edits or rejects it. It never sends on its own. Every decision is logged, so you can see what was approved and why a draft was rejected.
Why businesses buy it
Sales teams lose deals to slow follow-up, yet many are wary of fully automatic outreach that might send the wrong offer or the wrong tone to a valuable prospect. This agent removes the writing work and keeps your reps in charge of every word that goes out. Leads get a timely, relevant reply, and your team spends its time on judgment and conversations instead of drafting. The written scope covers the channel, two connected systems and up to six workflows before you order, so the boundaries are clear from the start.
Problems this solves
Reps spend hours writing follow-ups instead of talking to prospects
The agent drafts the outreach, follow-up or proposal that comes next, using your approved material. The rep you name approves, edits or rejects it, so the typing shrinks but the judgment stays theirs.
Fully automated sales emails feel too risky for your brand
Nothing goes to a prospect until a named rep signs off, and the agent never sends on its own. Your approval rules sit in up to six workflows, one per type of message.
Leads sit unqualified because nobody has time to sort them
The agent qualifies incoming leads in one channel, reading and updating up to two of your systems. It uses criteria from your documents, so reps see which leads deserve a draft first.
Managers cannot see why drafts were approved or turned down
Every decision is logged, showing what was approved and why a draft was rejected. That record helps you refine prompts over time and coach reps with real examples.
A good fit if
You want AI drafting sales conversations, but your team is not ready to let anything go out to a prospect unreviewed.
Your reps already qualify and respond to leads by hand and are looking for a faster first draft, not a fully autonomous agent.
You can name one rep or a small group who will approve, edit or reject drafts within an agreed response time.
You want a full history of what the AI proposed versus what a rep actually sent, for coaching and for accountability.
Not the right fit if
You want the agent qualifying and responding on its own, without a rep approving each message first. See the already-published AI Sales Agent, which holds full conversations directly.
The approval step you need is really about support replies, not sales outreach. See Human-in-the-Loop Support Agent, built around that workflow instead.
You need the queue and assignment logic for drafts waiting on a rep, not the sales agent producing them. See Agent Review Queue.
The travel-specific version of this pattern, quoting and booking with a person's sign-off, fits better. See AI Travel Sales Agent.
How it works
- A lead comes in through your chosen channel, and the agent reads it against your qualification criteria before deciding whether, and how, to respond.
- A reply, follow-up or proposal is drafted from your playbook: your pitch, your pricing and the objection responses you approved during setup, never improvised.
- The draft, with the lead's context and the agent's reasoning for that response, lands with the rep you named, inside the tool they already use.
- The rep approves it as written, edits it, or rejects it and tells the agent what to try instead; nothing reaches the prospect until one of those three happens.
- Approved messages send from your own CRM or channel, so the prospect sees a message from your team, sent on your system, not a third-party address.
- Every draft, edit and outcome is logged against the lead record, building a history you can use to refine the playbook or coach the team.
What you provide within 72 hours
Your portal lists these the moment you order. Delivery starts the working day after everything is in and your access is checked.
- Your business information (company basics, hours and brand voice), shared during intake.
- Admin or API access to your CRM and outreach channel, sandbox or test records if available.
- Your sales playbook: qualification criteria, pricing, and how you want objections handled.
- Ten real sample conversations or emails so drafts are built against genuine cases, not guesses.
- The rep or reps who will approve drafts, and how quickly they can turn them around.
- Admin access and API keys for every connected system, shared by adding our team as a user or creating a key you control, never by email.
- One decision-maker who can approve the playbook before launch.
- An API key for your own AI model account (for example OpenAI or Google Gemini), created for this build with a spending limit.
The 72-hour rule. Send everything on your checklist, including access, within 72 hours of ordering so your build starts on time.
Timeline
Start
Intake verified and your sales playbook agreed with you.
Weeks 1–2
CRM and outreach channel connected, drafting logic built.
Week 3
Approval routing configured for the rep or reps you named.
Weeks 3–4
Fifty documented test scenarios run.
Week 4
Your review and two rounds of changes.
Go-live
Live, recorded handoff session, support starts.
| Option | Delivery time | Price |
|---|---|---|
| Standard | 20–30 working days | $5,400 |
| Express | 15–25 working days | $6,200 |
| Rush | 10–20 working days | $7,000 |
What you receive at handover
- The sales agent live in one channel, with read and write access to up to two systems such as your CRM
- Up to six workflows carrying your approval rules, prompts and guardrails
- Your sales material loaded from up to a hundred documents
- A decision log showing what each rep approved, edited or rejected
- A test report covering fifty documented scenarios
- An admin guide, the recorded handoff session and 30 days of support
Expected benefits and how to measure them
- Reps can follow up with more leads without lowering the quality of each message.
- Helps keep every prospect message under human control.
- Makes it easier to learn which drafts work, using the decision log.
- Time to first follow-up
- Compare the gap from new lead to first message in your CRM for the month before and after launch.
- Drafts approved unchanged
- Use the decision log to count drafts approved without edits each week and watch the trend.
- Replies or meetings per rep
- Compare prospect replies or booked meetings per rep from your CRM, month before launch against month after.
Before you start: Pull your average time to first follow-up and replies per rep from the CRM. Ask reps roughly how long they spend writing messages each week.
These are things to track, not promised results. Outcomes depend on your data, your team and your market.
Running costs
The build price is one-time. Running Human-in-the-Loop Sales Agent afterwards costs roughly $50 to $250 a month in your own accounts for AI model usage and hosting, on top of whatever CRM plan you already pay for. AI usage during the build and testing is billed the same way, directly to your accounts, so every provider relationship stays yours.
Ask about running costsCustom requirements
If your playbook goes past the Scope Card, add a note at checkout. You get a written Change Quote, and that work only starts once you accept it. Your original order runs on its normal timeline either way. Larger or bespoke projects go through the Custom AI request form instead.
Request a custom versionKeep it running
We recommend monthly maintenance: monitoring, fixes and monthly change hours, which cover new objection responses or an updated qualification rule. Maintenance starts after the 30 days of support and can be cancelled any month.
$270/mo (5% of the build price a month, includes 2 change hours)
Optional. Starts after the post-launch support period. Cancel any month.
How maintenance worksHow this compares
| Do it yourself with a SaaS tool | Hire a custom agency | Upmark Labs | |
|---|---|---|---|
| Price | A sales engagement tool charges a recurring per-seat fee on top of AI add-ons | A custom sales-agent build is quoted after discovery | Ours is one fixed price, shown on this page before you buy |
| Scope | A generic AI writer leaves qualification and approval logic to you to configure | An agency agrees scope during the project | Ours ships with your playbook and approval rules already built in |
| Delivery | Engagement tools still need real playbook configuration before drafts are usable | An agency build usually takes several weeks, set per project | Ours has a fixed delivery time shown before you buy, with faster options |
| Control | Most AI sales tools send automatically once triggered | An agency deliverable depends on what was agreed | Ours never sends until your rep approves the draft |
Questions about this product
How much does Human-in-the-Loop Sales Agent cost?
The build is one fixed price, shown with each delivery speed in the panel on this page. That price covers the full Scope Card: one outreach channel, up to two systems such as your CRM with read and write access, and up to six sales workflows with your approval rules. AI usage and hosting afterwards are paid in your own accounts.
How is this different from a fully automatic AI Sales Agent?
The already-published AI Sales Agent qualifies and responds to leads directly, without a rep reviewing each message first. This build drafts the same kind of outreach but never sends on its own; the rep you name approves, edits or rejects every message before a prospect sees it. Choose this when your team is not ready to hand sending over entirely.
What does Human-in-the-Loop Sales Agent cost to run each month?
Running it afterwards costs roughly $50 to $250 a month in your own accounts for AI model usage and hosting, on top of whatever CRM plan you already pay for. There is no separate platform license on top of that figure. The exact cost depends on lead volume and how many drafts the agent produces.
What access does Human-in-the-Loop Sales Agent need to my systems?
We need admin or API access to your CRM and outreach channel, sandbox or test records if available, shared by adding our team as a user or creating a key you control, never by email. Your written playbook and ten real sample conversations or emails let us build qualification and drafting rules that reflect how your team actually sells.
Can we update the playbook or add workflows later?
Yes. Adjusting pricing, an objection response, or a qualification rule usually fits inside the monthly change hours of maintenance. Adding a genuinely new workflow beyond the six in scope, or a third connected system beyond the two included, goes through a Change Quote agreed before that work starts.
Can we cancel Human-in-the-Loop Sales Agent maintenance later?
Yes. Maintenance has no lock-in, so you can cancel any month once the 30 days of included support has ended. Unused hours do not roll over. Bigger changes to your playbook go through a Change Quote, priced before work starts.
Can the rep tell the agent why a draft was rejected?
Yes. Rejecting a draft includes a short reason, and the agent uses that reason on the next attempt for that lead rather than repeating the same message. Patterns in what gets rejected also show up in the logged history, which is useful evidence for refining the playbook during a monthly maintenance review.
Guides on this topic
All guides Let AI draft the outreach; keep your rep on the send button.
Order Human-in-the-Loop Sales Agent and every draft reaches a rep for approval before a prospect ever sees it, live on a fixed delivery date.
Works well with
Supervised Agents
Human-in-the-Loop Support Agent
Drafts support replies that an agent approves or edits before sending.