In short
Key takeaways
- Check what native workflow automation can do before you build anything in Make.com
- Make.com earns its place when you need branching, an AI model or a system outside Pipedrive
- Every recipe below writes a draft or a note. None of them sends email without a person reading it first
- The AI in Pipedrive takes one instruction per email. Make.com runs the same instruction on every deal
- Messy data stays messy. An AI model writes confident nonsense on top of empty fields
Source: Pipedrive and Make.com implementations by Sales Surge, Pipedrive features as of November 2026As of: 16 November 2026
In this article
- What can Pipedrive automate on its own?
- When do you need Make.com and an AI model?
- Worked example: from new lead to booked demo
- The 7 AI workflows, each in the same format
- In which order do you build this?
- Which tool is best for AI automation in Pipedrive?
- When should you not automate this?
- How we build it
- Glossary
- Frequently asked questions
Most teams that ask us about AI in Pipedrive want the same thing: fewer emails typed from scratch, notes that actually get logged and deals that do not go quiet without anyone noticing. Part of that is already in Pipedrive. The rest takes Make.com and an AI model. This guide starts with the native part, because building a Make scenario for something Pipedrive does with three clicks is money and maintenance you do not need.
What can Pipedrive automate on its own?
Pipedrive automates more than most users realise. Workflow automation runs on a trigger (a deal, person, organisation, activity or lead is created or updated) plus optional conditions, and then performs actions such as updating a field, creating an activity, sending a templated email or calling a webhook.
That covers what people usually mean by automated activities in Pipedrive. You set a rule like: when a deal enters the stage Proposal sent, create a call activity for the owner, due in three working days. Sequences go a step further and chain several activities and emails for one deal or person. Deal rotting marks a deal once it sits too long in a stage without activity.
Pipedrive also has its own AI. At the time of writing (November 2026) the AI email writer drafts an email from a short instruction you type, with a tone and a length you choose. It is available on Premium and higher plans. Pipedrive also summarises long email threads, can generate an AI summary of a deal, and the Sales Assistant suggests next steps. So does the AI in Pipedrive help reps with follow-ups and summaries? Yes. The difference with Make.com is precise: you type an instruction per email, but there is no fixed instruction set that runs on every deal, it does not read your call transcripts, and the result does not land in custom fields as structured data.
Which of these features you get, and how many automations can run at the same time, depends on your plan. Pipedrive changes this regularly, so check the current plan page before you decide.
| Native feature | What it does | Good enough when |
|---|---|---|
| Workflow automation | Trigger, conditions, actions inside Pipedrive | Trigger and result both live in Pipedrive |
| Activities per stage | Creates a task as soon as a deal moves | Your follow-up rhythm is the same for every deal |
| Sequences | A fixed series of activities and emails | The steps do not depend on what the prospect says |
| Deal rotting | Flags deals without activity for too long | You only need a signal, not a reason |
| AI email writer and summaries | Drafts an email from your instruction (Premium and higher) and summarises threads and deals | The rep is already in the mail and starts from there |
| Webhooks | Sends an event to another system | You want to start an external workflow |
When do you need Make.com and an AI model?
You need Make.com as soon as a workflow has to leave Pipedrive, branch on content, or call an AI model with your own instructions. Pipedrive starts the workflow with a trigger or webhook. Make fetches the context, sends it to the model, checks the answer and writes it back to the right deal.
We use five questions to decide. If you answer yes to one of them, native automation is not enough:
- Does the workflow need text that does not exist yet, such as a follow-up draft or a summary?
- Does it need data from outside Pipedrive, such as a call transcript, enrichment or your planner?
- Does the next step depend on what the deal or note says, not just on which stage it is in?
- Do you want the result as structured fields, not as a loose piece of text?
- Does it have to run on all open deals at once, for example every evening?
Not sure whether a workflow belongs in Pipedrive or in Make? We explain the difference on our Make.com page, including the questions we ask in an intake.
Worked example: from new lead to booked demo
This is the workflow people ask about most often: a new lead comes into Pipedrive, an AI model writes the follow-up email, the prospect books a demo through your planner, and the stage updates itself. Once the fields are in place, it is the quickest of the seven to build.
1. The lead comes in
A web form or chat writes the lead into Pipedrive with name, email, company and the question they asked. Make watches new leads, or Pipedrive calls a webhook.2. Make adds context
Make fetches the organisation and earlier notes. Optionally Apollo.io fills in company size and industry. Nothing is guessed: empty fields are passed on as empty.3. The model writes a draft
The model receives the lead data, your tone rules and the booking link of your planner, for example a Lumo booking page. It returns a subject, a body and a one-line reason why this angle.4. The draft lands in Pipedrive
Make saves the draft as a note on the lead with the label AI draft and creates an activity for the owner: review and send today.5. The prospect books
When the prospect books, the planner sends a webhook to Make. Make looks up the person by email, converts the lead into a deal if needed, moves it to Demo scheduled and creates the meeting activity.6. The team gets a signal
A short Slack message tells the owner what was booked, with a link to the deal.
The prompt we start from
You write a first follow-up email for a B2B lead.
Use only the data below. If a field is empty, do not guess.
Tone: plain, short sentences, no superlatives, address the reader as "you".
Maximum 120 words. End with one question and the booking link.
Lead: {{name}}, {{job_title}} at {{company}} ({{company_size}}, {{industry}})
Question from the form: {{message}}
Booking link: {{booking_url}}
Return JSON with the keys subject, body and reason.The 7 AI workflows, each in the same format
Each recipe lists the trigger, the steps, the fields you need, the Make.com modules, what it costs to run and the pitfall we see most. Costs depend on your own volume, so we count Make operations per run instead of quoting a price. The AI call costs between a fraction of a cent and a few cents per run, depending on the model and the length of the input.
Workflow 1: a follow-up draft after every meeting
The most useful one to start with, because every rep notices it the same day.
| Part | Detail |
|---|---|
| Trigger | An activity of type meeting, call or demo is marked done |
| Steps | Fetch the deal, notes and contact. Ask the model for a follow-up based on what was agreed. Save it as a note with the label AI draft. The rep reviews and sends |
| Fields needed | Activity note filled in, next step field, contact email |
| Make modules | Pipedrive Watch Activities, Get a Deal, List Notes, AI module, Create a Note |
| Running cost | About 5 operations per run, plus one AI call |
| Pitfall | Reps who write nothing in the activity note get a generic draft. Make the note a habit before you automate it |
Workflow 2: routing new deals to the right rep
Use native round robin rules first. Add AI only when the deciding factor is hidden in free text.
| Part | Detail |
|---|---|
| Trigger | A new deal or lead is created |
| Steps | Enrich the organisation. Let the model classify industry and size from the form text. Pick the owner from a fixed table. Post the reason in Slack |
| Fields needed | Industry, company size, region, a table of reps per segment |
| Make modules | Pipedrive Watch Deals, Apollo.io, AI module, Data store, Update a Deal, Slack |
| Running cost | About 6 operations per new deal |
| Pitfall | Letting the model choose the rep. It should classify; your own table decides. Otherwise nobody can explain an assignment |
Workflow 3: a one page brief before each meeting
Useful for teams with long sales cycles, where nobody remembers what was said three weeks ago.
| Part | Detail |
|---|---|
| Trigger | Scheduled check every 15 minutes for meetings starting within the next hour |
| Steps | Fetch the deal timeline, notes and open questions. Ask the model for a brief with three talking points. Save it as a note and send it to the owner |
| Fields needed | Meeting activities linked to a deal, notes with dates |
| Make modules | Scheduler, Pipedrive List Activities (filtered on type and date), Get a Deal, List Notes, AI module, Create a Note, Slack |
| Running cost | About 7 operations per meeting, plus the checks that find nothing |
| Pitfall | Pulling in web news makes the brief longer, not better. Start with your own CRM data only |
Workflow 4: a signal for deals that are slipping
Deal rotting tells you that a deal is stuck. This workflow adds why, and what to do next.
| Part | Detail |
|---|---|
| Trigger | Every evening at 18:00 |
| Steps | Fetch all open deals with their last activity and days in stage. Let the model rate the risk and give one suggested action. Create an activity on the high risk deals |
| Fields needed | Expected close date, deal value, last activity date, stage history |
| Make modules | Scheduler, Pipedrive Search Deals, Iterator, AI module, Router, Create an Activity, Slack |
| Running cost | Roughly 3 operations per open deal per evening, so this one grows with your pipeline |
| Pitfall | Scoring every deal every night with the model. Filter with plain rules first and only send the doubtful ones to the model |
Workflow 5: call summaries and action items
For teams calling through CloudTalk or another dialer that is linked to Pipedrive.
| Part | Detail |
|---|---|
| Trigger | A call is logged in Pipedrive with a recording or transcript |
| Steps | Fetch the transcript. Ask the model for a summary, decisions, objections and next steps as JSON. Save a structured note. Turn action items into activities |
| Fields needed | Calls linked to the right deal, a transcript available from the dialer |
| Make modules | Webhook or CloudTalk module, Pipedrive Get a Deal, AI module, Parse JSON, Create a Note, Create an Activity |
| Running cost | About 6 operations per call. Long calls make the AI call more expensive |
| Pitfall | Calls that are not linked to a deal end up nowhere. Fix the dialer link first. Tell callers that calls are recorded |
Workflow 6: a competitor brief at the proposal stage
Only worth it when your reps regularly lose to the same few competitors.
| Part | Detail |
|---|---|
| Trigger | A deal moves to the stage Proposal or Negotiation |
| Steps | Collect what the notes and emails say about alternatives. Let the model compare that to your own battle cards. Save a short brief on the deal |
| Fields needed | A competitor field on the deal, battle cards as text in a data store |
| Make modules | Pipedrive Watch Deals, List Notes, Data store, AI module, Create a Note |
| Running cost | About 5 operations per deal that reaches the stage |
| Pitfall | Asking the model what it knows about a competitor. It will make things up. Feed it your own battle cards and nothing else |
Workflow 7: restarting deals that went quiet
The re-engagement email is written for the specific deal, not copied from a template.
| Part | Detail |
|---|---|
| Trigger | A Pipedrive filter: open deals without activity for 14 days |
| Steps | Fetch the deal context and the last conversation. Ask the model for a short email that picks up where it stopped. Save it as a draft and alert the owner. No action within 48 hours goes to the manager |
| Fields needed | Last activity date, the reason the deal stalled if known, owner |
| Make modules | Scheduler, Pipedrive Search Deals, Iterator, AI module, Create a Note, Slack |
| Running cost | About 5 operations per stalled deal |
| Pitfall | Running it on deals that should simply be marked lost. Add a lost reason review before the email |
In which order do you build this?
Build one or two workflows, let them run for a month and look at what reps actually do with the drafts. Then add the next. Teams that switch on seven workflows in one week usually switch most of them off again, because nobody knows which one does what.
First
- Workflow 1 (follow-up drafts)
- Workflow 5 (call summaries)
Saves time every day and reps see the result straight away
Then
- Workflow 4 (slipping deals)
- Workflow 7 (quiet deals)
Needs a month of clean activity data to be useful
Later
- Workflow 2 (routing)
- Workflow 3 (meeting brief)
- Workflow 6 (competitor brief)
Depends on historical data and agreed rules per segment
Which tool is best for AI automation in Pipedrive?
There is no single best tool. Native automation wins whenever trigger and result both live in Pipedrive. Make.com is our default for anything with an AI model, branches or several systems, because a scenario shows every step visually and handles loops and error routes well. Zapier is fine for short, linear workflows and for teams that already use it everywhere.
Pricing models differ too: Make counts operations, Zapier counts tasks. With many steps per run, compare both on your own volume before you choose.
| Pipedrive native | Make.com | Zapier | |
|---|---|---|---|
| Best for | Rules inside Pipedrive | AI, branches, several systems | Short linear workflows |
| AI with your own prompt | Per email, by hand | Yes, on every run | Yes, on every run |
| Loops over all open deals | No | Yes | Limited |
| Error handling | Basic | Error routes per module | Basic retries |
| Who maintains it | Sales ops | Someone who knows Make | Sales ops |
When should you not automate this?
We say no to AI automation in Pipedrive more often than you might expect. These are the usual reasons:
- Your pipeline has no agreed stages yet. Then you are automating a process that does not exist.
- Activities and notes are rarely logged. The model has nothing to read.
- You want emails to go out without review. We do not build that, because one wrong email to a customer costs more than the time saved.
- You have a handful of deals a month. A good template does the same job.
Want to see how AI fits your whole CRM, not just single workflows? Read our AI-first CRM strategy. To score leads as well, see Pipedrive lead scoring with AI.
How we build it
Want this built for you? Our Pipedrive workflow automation service covers the native automations and the Make.com scenarios, including documentation your team can read. For a custom AI sales agent that reads and writes Pipedrive data, we run the AI Agent Sprint: from €4,997, built in 10 working days, with 30 days of support afterwards.
When an AI model needs direct access to CRM data, we use Surge Connect. It gives the model access to Pipedrive through an MCP server with OAuth 2.1, scoped per tenant, so you do not hand an AI tool a full admin token.
We are a Make.com Certified Partner and a Pipedrive Platinum Partner, so the scenarios and the Pipedrive side are built by the same team.
For more on follow-up, routing and tasks, see our sales automation guide.
Glossary
- Workflow automation
- The native automation in Pipedrive: a trigger, optional conditions and one or more actions.
- Make.com scenario
- A chain of modules in Make.com that starts on an event or a schedule and runs actions in other tools.
- Operation
- One module run in Make.com. Your Make plan is priced on the number of operations per month.
- Webhook
- A message one system sends to another when something happens, for example when a deal changes stage.
- Structured output
- An AI answer in a fixed format such as JSON, so Make can write each part to the right field.
- Deal rotting
- A native Pipedrive signal that marks a deal after a set number of days without activity in a stage.
Frequently asked questions
Does the AI in Pipedrive help reps with follow-ups and summaries?
Yes, within limits. At the time of writing, the AI email writer on Premium and higher drafts an email from a short instruction, with a tone and length you choose. Pipedrive also summarises email threads and deals, and the Sales Assistant suggests next steps. But you type the instruction per email: no fixed instruction set runs on every deal, call transcripts are not read and nothing lands in custom fields. For that you connect an AI model through Make.com.
Can Pipedrive create activities automatically?
Yes. With workflow automation you create an activity as soon as a deal enters a stage, a field changes or an activity is completed, with a due date a set number of days later. Sequences chain several activities and emails. Which options you have and how many automations can run at once depends on your plan, so check the current plan page.
Do I need Make.com, or is Zapier enough?
Zapier is enough for short, linear workflows: one trigger, two or three actions, no branches. Once a workflow calls an AI model, loops over all open deals or needs an error route, we use Make.com, because every step stays visible and you can handle each failure separately. Make counts operations and Zapier counts tasks, so compare both on your own volume.
Will the AI send emails without anyone checking?
Not to customers, in anything we build. Every workflow in this guide writes a draft or a note in Pipedrive and creates a task for the owner. The exception is internal output: a Slack message to the owner, a note on the deal or a filled field can run without review, because a wrong one costs a minute to correct. After a few weeks you refine the prompts, because the first drafts usually sound too generic.
Is my Pipedrive data safe when an AI model works with it?
It depends on how you set it up. Make connects to Pipedrive through the official API. Send only the fields a workflow needs, choose an AI provider with business terms and a data processing agreement, and check whether your data is used for training. When a model needs direct CRM access, Surge Connect gives it access through an MCP server with OAuth 2.1, scoped per tenant.
What does it cost to have AI automation built in Pipedrive?
Building workflows falls under our Sales Automation service, from €2,200 with a lead time of 2 to 3 weeks. A custom AI sales agent is the AI Agent Sprint, from €4,997, built in 10 working days with 30 days of support. On top of that you pay your own Pipedrive plan, a Make.com subscription based on operations, and the AI model per call.
Which workflow should I build first?
Start with the follow-up draft after a meeting or the call summary. Both save time every day, reps see the result straight away, and they need little history. A good sign to add the next one: reps send most drafts with only small edits. Routing and competitor briefs come later, because they only work once you have agreed rules and enough data per segment.
Citations & Sources
The following sources were referenced in creating this article. Links marked with require confirmation before publication.
- Pipedrive Knowledge Base: workflow automationhttps://support.pipedrive.com/en/article/workflow-automation
- Pipedrive API Documentationhttps://developers.pipedrive.com/docs/api/v1
- Make.com Pipedrive integrationhttps://www.make.com/en/integrations/pipedrive
- CloudTalk and Pipedrive integrationhttps://www.cloudtalk.io/integrations/pipedrive
- Pipedrive Marketplace: Sales Surgehttps://www.pipedrive.com/en/marketplace/partners/partner/sales-surge




