SynthesisLogic

AI agent for sales follow-up

An agent that watches your CRM and sales inbox for new leads, quotes without a reply and stalled deals, drafts follow-ups in your salespeople's voice, reads the replies and writes every touch back to the CRM.

Book an intro call How it works
A morning with this agent
  1. Found 7 quotes with no reply after 5 daysDone
  2. Drafted 7 follow-up emails for reviewAwaiting approval
  3. Updated 15 CRM records from yesterday's emailsDone
  4. Morning brief sent to 3 salespeopleDone

An example. Your agent is built around your own work.

How it works

  1. Find what needs a nudge

    It reads the CRM and the sales mailbox for new leads, quotes with no reply after the number of days you set, and deals that have gone quiet. It skips anyone who has replied, unsubscribed or been marked do-not-contact.

  2. Draft the message

    It writes a short follow-up from the last conversation and your salesperson's approved templates, tied to one reason to reply, such as the quote sent last week.

  3. Send or hold

    At first every draft waits for the salesperson to approve it. Once you trust a routine type of reminder, it can go from the salesperson's own mailbox within daily limits.

  4. Read replies and brief the team

    It sorts each reply as interested, not now, unsubscribe or out of office. Interested leads are offered meeting slots, everything else goes to the salesperson, and each morning every salesperson gets a short list of the deals that need a human.

New to AI agents? How they work, in plain words

Real examples, setup steps and tool connections for this agent

An example conversation

An example of how a conversation runs. The grey lines show what the agent does in your systems meanwhile.

  1. CustomerMorning. Anything I need to chase today?
  2. Read the CRM: 31 open deals for Omar. 6 quotes sent more than 5 days ago with no reply. Checked his mailbox: 1 of the 6 replied yesterday.
  3. AgentFive quotes need a nudge. Karim Trading is the oldest, quote 4411, sent 9 days ago. I have put a draft for each in your Outlook drafts folder. Want to see Karim's first?
  4. CustomerYes.
  5. Agent"Hi Karim, I wanted to check that quote 4411 for the packing line reached the right person. Happy to go through the numbers on a ten-minute call this week." It uses your usual sign-off.
  6. CustomerFine, send that one. Keep the others until I've read them.
  7. Sent the draft from Omar's mailbox. Logged the email on the deal and set a check for a reply in 3 days. Four drafts left waiting for approval.
  8. AgentSent. The other four are in your drafts.
  9. New reply from Delta Foods: "Please stop emailing me." Marked them do-not-contact in the CRM and cancelled 2 queued messages.
  10. AgentDelta Foods asked to be removed, so I've stopped all messages to them. One more: Nasser Co. replied to the last quote and the tone sounds unhappy. I have not answered it. That one needs you.

How it is built

It starts from your CRM
The agent reads leads, deals and activity from Salesforce, HubSpot, Pipedrive or another CRM through its API, or from a spreadsheet export if that is what you use. You set what counts as stale, for example no reply after five days for a quote.
It works in the salesperson's own mailbox
With each person's permission, mail is read and sent through the Gmail API or Microsoft Graph for work Outlook accounts. It checks for a reply before every message, so a sequence stops the moment the contact answers.
A language model writes, your rules decide
A large language model, the same kind of AI behind ChatGPT and Claude, writes the draft from the deal's history and your templates. Your rules decide who gets a message, when, and how many per day. Before every send it checks consent and your suppression list.
Approval and sending limits
A person approves the first messages of every new campaign. Daily sending caps apply, and sending pauses if emails start bouncing. For outreach we recommend a separate sending domain with SPF, DKIM and DMARC set up, so a problem cannot damage the reputation of your main company domain.
Replies are sorted, doubtful ones go to a person
The agent classifies each reply. It offers meeting slots from the salesperson's calendar only to clearly interested leads. If a reply is unclear, or sounds like a complaint, it goes to the salesperson untouched rather than being treated as interest.
A fixed list of actions
The agent can draft, send approved or routine messages, mark do-not-contact, offer calendar slots, update CRM fields and notify the team. It cannot change a price, add a discount, send a revised quote or mark a deal won or lost.
Logs, consent and where it runs
Every touch and outcome is written to the CRM, so you can see reply, meeting and unsubscribe rates per campaign. Consent rules differ by country (GDPR in the EU and UK, CAN-SPAM in the US), and opt-outs are honoured within the time the law allows. The system runs in the cloud or on your own servers.

What you can connect to it

Salesforce
Reads leads, contacts, tasks and opportunities, and writes activities and next steps.
HubSpot
Reads contacts, deals and engagement events, and logs every email and call outcome.
Pipedrive and other CRMs
Same reads and writes through each CRM's API, so the pipeline reflects what actually happened.
Gmail
Reads threads and sends approved follow-ups from the salesperson's own mailbox, with reply checks.
Outlook / Microsoft 365
Does the same for work accounts through Microsoft Graph, and puts drafts in the drafts folder.
Google Calendar and Outlook calendar
Offers meeting slots to interested leads and books the meeting.
WhatsApp Business Platform
Sends approved template follow-ups to contacts who agreed to hear from you and reads their replies.
Quoting tool or ERP
Reads quote numbers, amounts and dates so follow-ups quote the right details.
Slack or Microsoft Teams
Delivers each salesperson's morning list and takes questions from them in chat.
Google Sheets or Excel
Imports lead lists from trade shows or partners and keeps the pipeline if you have no CRM.
Book an intro call

Connects to your tools. You stay in control.

Works with what you have
Salesforce, HubSpot, Pipedrive and other CRMs, Gmail, Outlook / Microsoft 365, Google Calendar, WhatsApp Business Platform, your quoting tool or ERP, Slack or Teams, spreadsheets.
A person approves what matters
Payments, orders, discounts or anything you choose wait for approval. Every action is logged.
Can run on your own servers
The agent, its data, its logs and even the language model (open models such as Llama, Qwen or Mistral) can run on your servers, so documents and customer data stay inside your network. How that works

What it does not do

  • It does not email people who never contacted you or agreed to hear from you. Outreach to new prospects, if you want it, starts as drafts a person approves.
  • It does not change prices, offer discounts, send a revised quote or negotiate. Those go to the salesperson.
  • It does not mark a deal won or lost by itself.
  • It does not keep following up after a reply or an opt-out.
  • It does not promise pipeline or revenue results. What it changes in your numbers shows up only in your own data.

Questions

Does it send emails on its own?

Only if you allow it. Most teams start with drafts that a salesperson approves, then let routine reminders go out automatically within daily limits.

Will it sound like a robot?

It writes from your past emails and approved templates, and you review the first drafts until the tone is right. Short messages tied to one reason to reply read better than long ones, so that is what it writes.

Can it automate follow-up emails in Gmail or Outlook?

Yes. It finds threads with no reply after the days you set and sends the next step from the salesperson's own account, and the sequence stops when the contact replies. For Outlook work accounts it uses Microsoft Graph, not the consumer Outlook.com connector.

Why an agent instead of a Power Automate flow?

It depends on how simple the rule is. A flow that waits five days, checks for a reply and sends a fixed email is fine in Power Automate. An agent is worth it when each message must refer to the last conversation, replies need sorting, and the CRM should stay current without anyone typing.

We have no CRM. Is that a problem?

No. The agent can keep the pipeline and the log of touches in a spreadsheet, and we can help you move to a CRM later.

Could it hurt our email reputation?

Not if it is set up well. Sending caps, a pause on bounces and a separate outreach domain protect your main domain, and a person approves the first messages of each campaign. Sending from the main domain at high volume is what causes damage, so we avoid it.

Does it handle unsubscribes and consent?

Yes. It checks consent and the suppression list before every message, marks opt-outs in the CRM at once and cancels queued messages. Rules differ by country, so we set them up with you.

What if it misreads a reply?

Anything unclear, or anything that sounds like a complaint, goes to the salesperson without an answer. Only clear interest triggers meeting slots. Every classification is logged, so mistakes are visible and the rule behind them can be fixed.

How much does an agent like this cost?

It depends on the task and the systems it connects to. After a first call we give you a clear estimate of cost and time.

Want this agent built around your work?

Tell us in a few lines what the job looks like today and we will set up an online intro call.

Type of organisation