CMP42 // FAQ // CRMTERMS AND PROMPTS
// FAQ · CRM

CRM, WITHOUT THE SCREENS.

The familiar terms — lead, pipeline, lifecycle, forecast — and how the work gets done when the interface is the AI your team already uses.

This FAQ translates the classic CRM terms into the way a headless CRM works: the data sits in CMP42 in structured form, and the interface is the AI your team already uses, such as Claude, connected through the Model Context Protocol (MCP).

Every answer follows the same pattern: first what the term means, then how to put it into practice with CMP42 and an agent. The chapters follow the CRM glossary, from the basics to HubSpot-specific terms.

Where things stand: CMP42 has been in Alpha since October 2026 and works with design partners. Where a feature is only planned, the answer says so. Tool names and flows in the examples are illustrative and may still change during the Alpha.

Prompts with the Podma Connector: each question comes with up to three typical prompts you would give Claude with the Podma Connector connected. “Planned¹” next to a prompt means it relies on a feature that is not built yet. Three additions improve almost any prompt: “Answer only from Podma data”, “Name the records you relied on” and “Change nothing without my approval”.

Sources

Content as of: 1 November 2026

Product, pricing and Design Partner questions are in the product FAQ. Product FAQ

The chapters

Basics: CRM, headless and MCP

01
Basics: CRM, headless and MCP

What is a CRM, and what changes when AI becomes the interface?

A CRM (customer relationship management system) keeps track of contacts, companies, sales opportunities and every interaction with them. Classic CRMs such as HubSpot, Salesforce or Pipedrive are built for people who open forms, fill in fields and drag deals across a pipeline.

Many teams now ask their AI assistant about the pipeline first and only open the CRM when the answer looks odd. When the agent reads a data model that was designed for clicks, it is quick to invent fields or stages. CMP42 turns this around: the data model is built for agents, and the interface is the AI your team already works with.

Typical prompts with the Podma Connector

  1. Pipeline overview instead of a dashboard: Using the Podma Connector, give me an overview of all open deals by stage, with the count and total value per stage. Name the records you relied on.
  2. Spot stalled deals: Which deals have had no activity in the last 30 days? List owner, value and the latest note.
  3. Start of the day: What changed in my deals and contacts since yesterday, and which of it needs my attention today?
02
Basics: CRM, headless and MCP

What does "headless CRM" mean?

Headless means there is no user interface you are forced to use. There is no pipeline builder and no form for every field. Instead, CMP42 provides a structured graph of Context, Models and Processes and a fast API that any agent can read from and write to.

For keyboard work on large data sets there is the Podma Workbench: views, filters and saved views across a million records and more. Your team stays in its own tools, whether that is Claude, ChatGPT, Slack or in-house apps, and CMP42 supplies them with context.

Typical prompts with the Podma Connector

  1. Find your way around the workspace: Which sheets and object types exist in my Podma workspace, and how many records does each hold?
  2. A view without clicking: Create a view 'Q4 renewals' in the Workbench: deals in the renewal stage closing in Q4, sorted by value.
  3. An answer for the team channel: Summarise the five most important open deals so that I can post the text straight into our Slack channel.
03
Basics: CRM, headless and MCP

What is MCP, and how does Claude connect to CMP42?

The Model Context Protocol (MCP) is an open standard that lets AI applications use tools and data. An MCP server describes each tool with a name, a description and a typed input schema. The client, Claude for example, discovers these tools at runtime, and the model decides when to call them.

CMP42 is MCP-native: every Model automatically becomes a set of MCP tools. You point Claude at your workspace, and it can read, write and plan against your real context. In practice:

  • Clients: Claude Desktop, Claude Code, Cursor, Cline and your own agents built on the MCP SDK.
  • Transport: local stdio for development and the CLI, or Streamable HTTP for hosted agents.
  • Access: a workspace-scoped API token in the MCP client configuration, limited to one workspace and named Models. OAuth 2.1 follows after GA.
  • Without MCP: Slack bots, internal apps and scripts use REST or GraphQL, with the same authentication and the same audit trail.

Typical prompts with the Podma Connector

  1. Check the connection: Check the connection to the Podma Connector: which workspace can you access, and which tools are available to you?
  2. Reading versus writing: Which of your Podma tools only read, and which can change data? Sort them into two lists.
  3. A first harmless test: As a harmless test query, show me the five most recently changed deals with date and owner.
04
Basics: CRM, headless and MCP

How are contact, company, lead, deal and activity represented in CMP42?

CMP42 ships with lean standard Models: contact, company, deal, task and activity. They match the basic terms of the glossary. The difference is in the connections: relationships are typed edges. A contact knows a company, owns a deal, authored a document.

In CMP42 a lead is a stage in a contact’s lifecycle, not a separate record that has to be “converted” later. Seven fixed lifecycle stages, the same for every tenant, are planned for the next version: captured, in_qualification, qualified, opportunity, customer, disqualified and transient. Every record is versioned, and every field can be addressed by URI.

Typical prompts with the Podma Connector

  1. Create without a duplicate: Create the contact Anna Weber (anna.weber@example.com) and link her to Example Ltd. Check first whether she or the company already exists.
  2. Show the relationships: Show me all contacts at Aspen Ltd with their role and the deals they are involved in.
  3. Gaps in the links: Which contacts are not linked to any company? Suggest links based on the e-mail domain, but do not change anything.
Decided1
05
Basics: CRM, headless and MCP

What happens to activities such as calls, e-mails and meetings?

Activities and notes are records in the graph, linked to the contact, the company and the deal. This includes content that classic CRMs leave out: documents, call recordings, product briefings and playbooks.

In practice, Claude can write a note back after a call with the write_note tool. The note is validated, versioned and attributed to the agent that wrote it. In the next conversation the agent reads it again.

Typical prompts with the Podma Connector

  1. Wrap up a call: Summarise this call transcript in five points and save it as a note on the deal 'Aspen Q4 renewal'. [paste transcript]
  2. Read back the conversation history: What did we discuss with Example Ltd over the last three months? In chronological order, with the source of each statement.
    Open - you decide1Customer journey timeline
  3. Record the next step: Create a task for me: follow up on the Aspen proposal by Friday, linked to the deal.
Decided1
06
Basics: CRM, headless and MCP

Why not just paste a CRM export into ChatGPT, or use Notion with Claude?

An export is a snapshot without structure, and Notion is a collection of documents. An agent answers questions from them as it would in a chat, and has to guess. In CMP42, asked “Who at Aspen owns the Q4 renewal?”, the agent follows named edges (contact owns deal) and names the records it read. The answer can be checked instead of merely sounding plausible.

Typical prompts with the Podma Connector

  1. An answer with evidence, not a guess: Answer only from Podma data: who owns the Q4 renewal at Aspen? If the data does not say, tell me so explicitly.
  2. Reconcile an old export: Compare the attached export file with the records in Podma and list differences in stage, owner and value.

Sales and lead management with agents

01
Sales and lead management with agents

How do new leads get into CMP42, and what does the agent do?

CMP42 does not collect leads by itself: there is no tracker and no pixel. Leads arrive by import or through the API; in the Alpha, for example, by CSV import for contacts, companies, deals and notes. Lead forms from Meta, Google and LinkedIn are planned to come in labelled with what each of them proves in legal terms.

The agent takes over once a lead has arrived: it checks whether the person is already known, links them to their company and proposes the next step. Every contact carries its legal basis, consent date and consent source (see chapter 3).

Typical prompts with the Podma Connector

  1. Pre-check a trade-fair list: Check the attached CSV of trade-fair contacts: who already exists in Podma, who is new, and who lacks a legal basis? Do not change anything yet.
  2. HubSpot import: Import our HubSpot contacts into a sheet and show me the summary: new, changed, no longer in HubSpot.
  3. Review form submissions: Which leads came in through our website form this week, and do all of them have confirmed consent?
Decided1
02
Sales and lead management with agents

How does an agent qualify leads with CMP42?

The principle is: read first, then ask. Classic qualification flows ask even a three-year customer what their company does. With CMP42 the agent reads the context before the first question: is the person known, which company do they belong to, is there an open deal, what was agreed last time?

A typical flow:

  1. A request comes in, and the address matches an existing contact.
  2. Over MCP, the agent reads the company, the open deal and the note from the last call.
  3. It skips the questions that have already been answered and asks only the open one.
  4. The answer is written back as a qualification record, linked to contact, company and deal.
  5. A Process puts it in front of the deal owner for approval.

Note that CMP42 is the layer the agent reads, not the chatbot itself. There is no widget and no messaging integration; your channel stays yours. Structured assessments, BANT for people and MEDDICC for deals, are planned for the next version.

Typical prompts with the Podma Connector

  1. Read first, then ask: Before I talk to Max Mustermann: is he known to us? Which company, which open deal, what was agreed last time? Which qualification question is still open?
  2. BANT with evidence: Assess this lead against BANT using the data we have, and mark which criteria are backed by evidence and which are still open.Decided1BANT and MEDDICC
  3. Leads left lying: List every lead in qualification that nobody has spoken to for 14 days, sorted by company size.
Decided1
03
Sales and lead management with agents

What is the difference between an MQL and an SQL, and how do you model it?

An MQL has shown interest that marketing considers relevant. An SQL has been confirmed by sales as a real sales opportunity. Classic CRMs usually compute the transition with a points score that nobody can explain afterwards.

CMP42 deliberately ships no built-in score and no ready-made MQL. Instead it stores every buying signal as a record of its own, with type, source system and time. What counts as buying intent is set in your Model. When sales asks “Why this lead?”, the agent answers with the specific signals behind it, dated and with their source, rather than with a number.

Typical prompts with the Podma Connector

  1. Missed buying intent: Which contacts replied or booked a meeting in the last 30 days but are not yet with sales as an SQL?
  2. "Why this lead?": Why is Example Ltd considered a hot lead? Name the specific signals with date and source.
    Open - you decide1Buying signals as objects
  3. Derive an MQL threshold: What did the leads that became won deals last year have in common? Suggest criteria for an MQL threshold from that.
Decided1
04
Sales and lead management with agents

How does a sales funnel differ from a sales pipeline?

The funnel is a model: it describes how many people reach each step on the way from first contact to purchase. The pipeline is concrete: it holds the real deals and their current stage.

In CMP42 the deal stages are declared in your Model, and tools such as list_deals_at_stage are generated from them. If an agent sends a stage that does not exist, it gets a clear error and corrects itself instead of inventing a stage. You build the funnel by having the agent count and compare the contacts’ lifecycle stages.

Typical prompts with the Podma Connector

  1. Read the pipeline: Show me the pipeline: number and total value of deals per stage, with the largest deal in each stage.
  2. Read the funnel: How many contacts are in each lifecycle stage, and where is the biggest drop between two stages?Decided1Seven lifecycle stages
  3. Clean stages: Which deal stages are defined in our Model, and are there deals with stage values outside that list?
Decided1
05
Sales and lead management with agents

How do you calculate conversion rate and win rate with an agent?

The formulas stay the classic ones:

  • Conversion rate = the number who took an action / the number who could have taken it.
  • Win rate = won deals / all closed deals (won plus lost).

CMP42 has no built-in insights layer for this. The agent reads the deals with their stages and versions and does the arithmetic itself. Because every change is versioned, it can also say when a deal reached each stage. It can name the records a figure rests on, so every number can be traced back to the record.

Typical prompts with the Podma Connector

  1. Win rate with evidence: Calculate the win rate for Q3: won / (won + lost) deals. List the deals you counted.
  2. Conversion by source: What was the lead-to-deal conversion per source (form, trade fair, referral) over the last six months?
  3. Find the bottleneck: In which stage do deals stay longest, measured by the median time in stage?
    Open - you decide1Funnel metrics over MCP
Decided1
06
Sales and lead management with agents

How is a forecast produced?

A forecast predicts revenue from open and expected deals. With CMP42 you ask your agent, for example: “Which deals are likely to close in Q4, and which are at risk?” The agent reads deal values, stages, recent activities and notes, and justifies its estimate with the records it read.

The judgement stays with people. A Process can put the forecast up for approval, and the decision is recorded with its author. If the Model or the flow changes, the forecast can be rerun against an earlier state of the data and compared.

Typical prompts with the Podma Connector

  1. Weighted forecast: Build a Q4 forecast from all open deals expected to close in Q4. Weight by stage and justify each estimate with the latest activity.
  2. Risks in the forecast: Which deals in the Q4 forecast are at risk because the contact person changed or nothing has happened for three weeks?
  3. Forecast against actuals: Compare the approved forecast of 1 October with today's state and explain the largest differences.
    Open - you decide1Approved forecast snapshots
Decided1

Marketing and customer retention

01
Marketing and customer retention

Can CMP42 replace marketing automation?

Not entirely, and that is deliberate. Sequences, sending and opt-out management are not part of the Alpha. CMP42 takes the part that comes before them, the part classic tools often reach too late: it checks who may be contacted at all and why, and it rejects a recipient list that breaks the rules.

The work is split like this: the agent proposes segments, content and the next step. CMP42 supplies the context and checks the rules. A person approves. Your existing tool still does the sending.

Typical prompts with the Podma Connector

  1. Audience with an eligibility check: Build an audience: managing directors of agencies in the DACH region with no proposal in the last six months. How many of them may be e-mailed, and why are the others excluded?
  2. Draft for the mail tool: Draft an invitation e-mail for this audience in the tone of our last campaign. I will approve it and send it through our mail tool.
02
Marketing and customer retention

What is the difference between a workflow and a CMP42 Process?

A classic workflow is triggered by an event and runs actions automatically, often out of sight. A Process in CMP42 is a named sequence of steps, some carried out by an agent and some by a person. Every step reads and writes the same graph, and every transition is logged.

What it adds over a Zapier hop:

  • Approval steps (human in the loop): the Process waits until a person agrees. Approval can be given in the Podma Workbench, through an MCP tool, through a signed link by e-mail, or through a Slack or Teams bot.
  • Run history and replay: every run can be traced step by step and rerun against an earlier state of the data.
  • Cost transparency: token usage is visible per Process and per run, in euros as well.

How exactly Processes are declared is being designed by protagx together with the design partners during the Alpha.

Typical prompts with the Podma Connector

  1. Start a Process: Start the renewal-note flow for the Aspen deal: read the context, write a draft, put it in front of me for approval.Decided1Process runtime
  2. Check runs and cost: Show me the last ten runs of the renewal-note Process with status, approval and token cost.Decided1Runs with cost
  3. Translate a workflow: When a deal moves to 'won', a project should be created automatically and put in front of the owner for review. Draft the Process definition.
    Open - you decide1Event-driven processes
Decided1
03
Marketing and customer retention

How does lead nurturing work with an agent?

Nurturing means keeping a relationship alive with relevant content until the contact is ready to buy. The weak point of generic AI text is that it does not know the history. With CMP42 the agent reads the account before every draft: who owns the deal, which documents exist, what the latest notes say.

The effect grows over time. Every approved note becomes part of the graph that the next run reads, so the second message starts with more knowledge than the first.

Typical prompts with the Podma Connector

  1. Follow-up from the real history: Draft a follow-up e-mail to Max Mustermann that picks up from our last conversation. Use only what is in Podma, and name your sources.
  2. Prospects gone quiet: Which contacts requested a guide, have heard nothing from us for 60 days and may be contacted by e-mail?
  3. Write the knowledge back: Save the approved e-mail as a note on the contact so that the next draft builds on it.
04
Marketing and customer retention

How do the customer journey and the customer lifecycle differ?

The customer journey describes everything a customer experiences across all touchpoints. The customer lifecycle describes the phases of the relationship, from acquisition to retention.

In CMP42 the journey appears as a sequence of events and signals in the graph, each with its source and time. The lifecycle is a stage on the contact; the plan is for every change of stage to state its reason. An agent can then answer both questions: “What has this customer been through?” and “Where do they stand right now?”

Typical prompts with the Podma Connector

  1. Tell the journey: Tell me Example Ltd's customer journey: first contact, forms, meetings, proposals, close, each with date and source.
    Open - you decide1Customer journey timeline
  2. Lifecycle status: Which lifecycle stage is Example Ltd in, since when, and what triggered the last change?Decided1Seven lifecycle stages
Decided1
05
Marketing and customer retention

How does CMP42 support customer retention?

Retention is often decided at the renewal. A typical use case for product-led B2B teams: an agent reads the real customer history and drafts the renewal note, and a partner reviews and approves it. If a note points to a risk, such as a new budget holder, it is recorded on the company with its author.

Not to be confused: “retention” also means the retention period for data. CMP42 has an answer for that too, the project record after the close (see chapter 6).

Typical prompts with the Podma Connector

  1. Renewals in view: Which customers' contracts end in the next 90 days? Sort by revenue and flag risks from the latest notes.
  2. Renewal note: Draft the renewal note for Aspen from the real history (usage, open issues, contact people) and put it in front of me for approval.
  3. Spot silence: Which existing customers have had no documented contact with us for six months?
06
Marketing and customer retention

How does an agent find cross-selling and upselling opportunities?

Cross-selling sells complementary products, upselling a higher-value solution. CMP42 does not deliver ready-made recommendations for either. But the agent can read what it needs for a good recommendation: closed deals and projects, the people involved and their roles, notes on needs and budget.

An example prompt: “Which existing customers bought product A in the last 12 months but never discussed product B? Justify each suggestion.” Before anyone is contacted, the same question applies as everywhere else: may this person be contacted on this channel?

Typical prompts with the Podma Connector

  1. Cross-selling candidates: Which existing customers bought product A but never discussed product B? Justify each suggestion.
    Open - you decide1Products and line items on deals
  2. Upselling signals: Which customers are on the smallest plan and have mentioned growth, new locations or more users in notes?
  3. Prepare the approach: For the three best candidates: who is the right contact person, and may I contact them by e-mail?
Decided1
07
Marketing and customer retention

How does CMP42 make sure we contact people lawfully?

CMP42 asks two questions in a fixed order. First: may anyone contact this person at all, on which channel, in which country and for which purpose? Second: is the contact worth making? That way a campaign does not fail at the send button after four weeks of work.

Already in the Alpha (since October 2026) Planned for the next version
Legal basis, consent date and consent source on every contact Reachability check per channel, country and purpose, with a reason
No recipient list without an eligibility check, 1 million records measured in 2.1 s Live count of eligible recipients while you define an audience
Eligibility check at any time on any view Jurisdiction templates, starting with the EU/EEA, Switzerland and the UK

CMP42 does not decide for you whether a campaign is lawful. You remain the controller; the templates support your documentation under GDPR Art. 5(2) and Art. 7(1) and are not legal advice.

Typical prompts with the Podma Connector

  1. Check a list before sending: Check the view 'Newsletter November': how many contacts may be e-mailed, and for what reasons are the others excluded?
  2. Documentation gaps: Which contacts have a legal basis but no consent date or no source? List them for review.
  3. A single case: May I invite Anna Weber to our webinar by e-mail? Justify it with channel, country and legal basis.Decided1Reachability check
Decided1

Data, Models and technology

01
Data, Models and technology

What are CRM objects, properties and associations called in CMP42?

The concepts stay; the implementation becomes stricter and readable for agents:

Glossary term Classic CRM CMP42
CRM object contacts, companies, deals, custom objects Model, declared in a file, versioned in Git
Property / field a field on a settings page typed field in the Model, validated on write
Association a link, often without meaning typed edge with a meaning (knows, owns, authored) and a validity period
Activity / note an entry on a timeline versioned record with an author (person or agent)

Typical prompts with the Podma Connector

  1. Understand the fields: Which fields does the 'deal' Model have, which are required, and which values does the stage field allow?
  2. Show the links: Show me all of Max Mustermann's links: companies, deals, notes and documents, each with the type of relationship.
02
Data, Models and technology

How do Models work, and what do they have to do with MCP tools?

Models are your ontology: the list of business objects, their fields and their relationships. They are declared in a TypeScript-style YAML format, versioned in Git, reviewed like code and migrated with a single command.

On every migration CMP42 generates the matching MCP tools. Each Model gets six standard tools following a fixed pattern: create, read, update, delete, list and search, for example contact.create or deal.update. Models with tagged fields also get filter. Add a Model Engagement for consulting mandates, and Claude can list mandates by stage and write notes on them straight away, without anyone writing code.

Migrations are backward compatible, so tools that agents rely on in production keep working.

Typical prompts with the Podma Connector

  1. Design a Model: We need an 'Engagement' object for consulting mandates with title, stage, daily rate and responsible partner. Draft the Model file.
  2. Check the new tools: Which tools were generated for the 'Engagement' Model after the migration, and what does each one do?
  3. Use the new Model: List all engagements in the 'active' stage with responsible partner and daily rate.
03
Data, Models and technology

How does data quality stay high when agents write?

Data quality means complete, correct, current and consistent. Agents are both a risk and an opportunity here. CMP42 works on four points:

  • Validation on write: if a value does not fit the Model, the agent gets a clear error and corrects itself.
  • Lean standards: no contact with 200 fields. What is not in your ontology does not exist.
  • Attribution: every change names a person or an agent as its author.
  • Versioning: every record has a history with time and source.

Typical prompts with the Podma Connector

  1. Find duplicates: Find possible duplicate companies (same domain, similar name) and suggest merges without changing anything.Decided1Duplicate clean-up
  2. Gaps per field: Which contacts have no e-mail, no company or no legal basis? Give me a separate list with a count for each gap.
  3. Review agent changes: Which records did an agent change last week? List record, field, old and new value.
    Open - you decide1Record history over MCP
Decided1
04
Data, Models and technology

How does a data migration to CMP42 work?

In four steps:

  1. Inventory: list the objects, fields and relationships your team and your agents actually use. That is the first draft of your ontology.
  2. Mapping to Models: contacts, companies and deals land on the standard Models. Whatever survives the inventory is declared explicitly. Relationships hidden in free-text fields become typed edges.
  3. Import: because every write is validated, bad data shows up as clear errors instead of slipping quietly into the graph.
  4. Connect the agents: Claude and other MCP clients use the generated tools; the rest of your stack uses REST or GraphQL.

The Alpha has a CSV import for contacts, companies, deals and notes. Custom fields, histories and attachments are migrated by the founder at no extra cost as part of the design-partner programme. A HubSpot connector with OAuth arrives with the Beta in the first quarter of 2027.

Typical prompts with the Podma Connector

  1. Inventory: Import our HubSpot contacts into a sheet and show me which properties are never or almost never filled.
  2. Mapping proposal: Build a mapping proposal from the HubSpot properties to the Podma standard Models and mark what we can leave out.
  3. Merge under control: Merge the 'Industry' column from the import sheet into the Companies system sheet, matched on the HubSpot ID. Show me the preview first, with the count and the unmatched records.
05
Data, Models and technology

What is the difference between an API, an integration and MCP?

  • An API is a programming interface. CMP42 offers REST and GraphQL, meant for programs that developers write.
  • An integration is a specific connection between two systems, such as a sync between two CRMs. It has to be built and maintained.
  • MCP is an interface for AI models: the model discovers the tools itself and decides when to use them. A new AI client does not need a new integration project.

All three routes use the same authentication and the same audit trail. MCP is not a back door.

Typical prompts with the Podma Connector

  1. Pick the right route: Which of my recurring tasks can you do yourself through the Podma Connector, and which would need a script against the REST API?
  2. From conversation to script: Write me a short script that uses the REST API every Monday to fetch all deals in the renewal stage and save them as CSV.
06
Data, Models and technology

Is CMP42 the "single source of truth"?

CMP42 is meant as the central context layer that every agent reads. It does not promise that contradictions disappear. When two sources disagree, for example two CRMs with different lifecycle stages for the same customer, the graph shows both values with their origin and timestamp.

The decision about which value applies is made by a person, and the Process records it with its author. So “single source of truth” here does not mean “one value”. It means “one place where you can see why a value applies”.

Typical prompts with the Podma Connector

  1. Find contradictions: For which companies do the lifecycle stages from HubSpot and Podma contradict each other? Show both values with timestamps.
    Open - you decide1Source records with provenance
  2. Record a decision: For company Y the stage from the sales CRM applies. Record my decision with the reason.
    Open - you decide1Source records with provenance
Decided1

HubSpot terms and the way to CMP42

01
HubSpot terms and the way to CMP42

What corresponds to a HubSpot custom object, and when do you need one instead of a property?

In HubSpot a custom object is a user-defined object type and a property is an additional field. In CMP42 both are declarations in the Model. The rule of thumb stays the same: if something has fields of its own, relationships of its own and a lifecycle of its own, it is a Model of its own. If it is only an attribute of an existing object, it is a field.

An example from CMP42 itself: what applies after the close (contract, data processing agreement, data categories, retention period) is not three extra fields on the deal but a project record of its own, with its own edges to the deal, the contracting party and the people involved.

Typical prompts with the Podma Connector

  1. Model or field: In HubSpot we maintain a custom object 'Locations'. Should it become its own Model in Podma or a field on the company? Justify it from the fields and relationships.
  2. Find legacy clutter: Which custom properties on the imported HubSpot contacts are more than 80 percent empty?
02
HubSpot terms and the way to CMP42

What is the difference between lifecycle stage, lead status and deal stage?

The three terms describe different perspectives and should never be used as synonyms. CMP42 keeps them apart on purpose:

Term The question it answers Belongs to In CMP42
Lifecycle stage Where does the relationship stand overall? contact, company seven fixed stages for every tenant planned, each change with a reason
Lead status How far has work on this lead progressed? contact in sales result of a qualification record, e.g. qualified, nurture, rejected
Deal stage Which phase is this sales opportunity in? deal stages declared in the deal Model, validated on write

The benefit for agents: because every stage is typed, Claude cannot invent one. Ask “Which stage is the deal in, and where is the contact in the lifecycle?” and you get two separate answers, each with its source.

Typical prompts with the Podma Connector

  1. Inconsistent customers: List contacts with the lifecycle stage 'customer' that have no won deal.
  2. Contradictory statuses: Find contacts with the lead status 'rejected' whose company still has an open deal.
  3. Keep three perspectives apart: For Example Ltd, explain separately: the company's lifecycle stage, the contact person's lead status and the stage of the open deal.
Decided1
03
HubSpot terms and the way to CMP42

What corresponds to a HubSpot workflow and a HubSpot sequence?

A HubSpot workflow automates CRM processes through triggers and actions. Its counterpart is the CMP42 Process: steps by agents and people, with approvals, run history and cost per run (see chapter 3).

A HubSpot sequence is a series of e-mails and tasks for personal follow-up. CMP42 has no counterpart for it in the Alpha, because it does not send. A sensible pattern: the agent drafts each message from the real contact history, a person approves it, and your existing tool sends it.

Typical prompts with the Podma Connector

  1. Carry over a workflow: Here is the description of our HubSpot lead-assignment workflow. Translate it into a Podma Process with an approval step. [paste description]
  2. Sequence by hand: Draft three follow-up e-mails to Max Mustermann for day 4, 10 and 21, each from his history. I will approve them one by one.
    Automated:Open - you decide1Follow-up sequences
Decided1
04
HubSpot terms and the way to CMP42

What corresponds to dynamic and static HubSpot lists?

A dynamic list updates itself from criteria. In CMP42 that is a saved view with filters in the Podma Workbench, or a query the agent reruns whenever needed. A static list is a snapshot. Its closest counterpart is a recipient list that has been checked against your rules and approved.

Typical prompts with the Podma Connector

  1. Dynamic list: Create a saved view: all contacts in Berlin with the role managing director and documented e-mail consent.
  2. Static list: Freeze the view 'Webinar invitation' as a recipient list, with eligibility status and reason for each contact.
    Open - you decide1Frozen audiences
Decided1
05
HubSpot terms and the way to CMP42

We use HubSpot for marketing and a second CRM for sales. Does CMP42 help?

This is one of the typical starting points. A two-way sync moves values but does not decide which one is right. CMP42 places both records as objects next to the same company, each with its source system and timestamp. Before the customer meeting, the account manager asks her agent which stage applies. The agent shows both values with their origin, she decides, and the Process records the decision. Neither CRM was opened.

Both systems keep running; CMP42 does not require a migration first. In the Alpha the data comes in by import; there is no live connector yet.

Typical prompts with the Podma Connector

  1. Differences between systems: Which companies exist in both systems with a different stage? Show both values and the time of the last change.
    Open - you decide1Source records with provenance
  2. Meeting preparation: Prepare me for the meeting with company Y and point out contradictions between the marketing and the sales data.
Decided1
06
HubSpot terms and the way to CMP42

Who is a good fit for moving from HubSpot to CMP42, and who is not yet?

A good fit Not a fit (yet)
Your team mainly reaches CRM data through agents, or wants to Your team wants a click CRM with drag-and-drop pipelines and forms
You want your ontology in Git, reviewed like code You need marketing automation with sending today
You have developers or a partner who can work with an API You need a production system today; the Alpha ships weekly releases
You need EU hosting or self-hosting You cannot run a real workload in the Alpha; then the Beta from Q1 2027 is the right time

Typical prompts with the Podma Connector

  1. Fit check from data: Analyse our HubSpot import: how many object types, properties and records do we actually use? Does a lean headless CRM suit us?
  2. Name the gaps: Which features we use in HubSpot have no counterpart in Podma today?

Trust, data protection and control

01
Trust, data protection and control

Can we trace what an agent did in the CRM?

Yes. Every change records whether a person or an agent made it, and every record is versioned. When an agent runs inside a Process, the complete run history, the approval steps and the token usage are added. A run can be repeated against an earlier state of the data. That answers the question: was the error in the data or in the agent?

The comparison with Git fits: every change is a commit with an author, the approval is the review, and the run history is the log.

Typical prompts with the Podma Connector

  1. Who changed what: Who last changed the stage of the Aspen deal, a person or an agent, and what was there before?
    Open - you decide1Record history over MCP
  2. Review an agent's day: Show me all writes by agents from yesterday, grouped by object type.
    Open - you decide1Record history over MCP
  3. Narrow down an error: Rerun the renewal-note run of 15 October against the data as of 30 September and show the differences.Decided1Replay against an earlier state
Decided1
02
Trust, data protection and control

What may an agent do, and what not?

Access runs through an API token that is limited to one workspace and named Models. Agents follow the same rules as any other API user. In the Podma Workbench, agents may rearrange views and columns by command; commands that write or delete values are blocked for them.

At critical points you add approval steps: whatever an agent drafts moves on only after a person agrees. It is also planned that CMP42 refuses to delete a record while a legal hold is set or the retention period has not ended.

Typical prompts with the Podma Connector

  1. Show your permissions: Which permissions does the token you are connected with have? Which Models may you read, and which may you write?
  2. Test the limits without risk: Which delete or write tools are available to you? Do not call any of them; just list them.
  3. Limit field access: For this analysis, work only with company data, not with contact data or notes.
    Technically enforced:Open - you decide1Field-level access per MCP token
Decided1
03
Trust, data protection and control

Where is our data, and what leaves our environment?

  • Managed option: hosted in the EU. No data leaves the environment except what you explicitly send to a connected agent.
  • Self-hosting: the same software and the same Models on your own Postgres database and object storage. The image arrives with the Beta in the first quarter of 2027.
  • No third-party analytics, no fingerprinting, no session replay.

One point to decide deliberately: whatever Claude or another agent reads from CMP42 goes to the provider of that model. You control which data an agent may see through the token’s permissions; review your AI provider’s terms separately.

Typical prompts with the Podma Connector

  1. Make the data flow visible: Which data did you read from Podma in this conversation? List object types and fields.
  2. Ask sparingly: Answer my question with aggregate numbers only, without showing names or e-mail addresses.
04
Trust, data protection and control

Does CMP42 train AI models on our data?

No. CMP42 does not train on your data and does not ship a model of its own that could. You bring the model you trust.

05
Trust, data protection and control

What about compliance and certifications?

CMP42 follows the principle of data protection by design. A data processing agreement (DPA) is available on request. SOC 2 Type II is planned for general availability in the third quarter of 2027. The Alpha runs with the same internal controls but has not yet been externally audited or certified.

Typical prompts with the Podma Connector

  1. Overview for the DPO: Prepare an overview for our data protection officer: which object types and fields in our workspace hold personal data?
  2. Check legal bases: How many contacts have no documented legal basis, broken down by origin?
Decided1
06
Trust, data protection and control

What happens to customer data after the close?

Classic CRMs stop at “closed won”. In CMP42 the deal becomes a project that carries what applies legally from then on: contract reference and term, a reference to the data processing agreement (GDPR Art. 28), data categories with special categories shown separately (GDPR Art. 9), retention rule, retention date and legal hold.

The auditor’s question “On what basis do you hold this data, and until when?” becomes a query. Claude can ask, for example: “Which projects are past their retention date?” Contracts and files stay where your legal team keeps them; CMP42 stores references and does not remind you of deadlines automatically.

Typical prompts with the Podma Connector

  1. Expired retention: Which projects are past their retention date and have no legal hold?
  2. Special categories: Which projects contain Art. 9 data, and under which DPA do they run?
  3. Prepare for deadlines: Which contracts end in the next 60 days? Create a review task for each.
    Automatic:Open - you decide1Reminders before retention ends
Decided1
07
Trust, data protection and control

Is CMP42 open source?

No. CMP42 is proprietary software by protagx. Source code access for security reviews, audits and escrow is available on request under NDA.

Getting started

01
Getting started

Which teams is CMP42 meant for first?

For teams whose way of working has already moved to AI while their CRM has not. Four groups are in focus:

  • Boutique consultancies: your own framework becomes context the agent uses in the next engagement, instead of a slide copied into every deck.
  • Product-led B2B teams: agents draft renewal and follow-up notes from the real customer history.
  • Agencies with AI ambitions: you build agents for your clients that work on their real accounts instead of hallucinating.
  • Teams leaving HubSpot: you want lean data and a real API, not the next platform.

Typical prompts with the Podma Connector

  1. Consulting: Read our playbook for first meetings and the history of client X, and draft the agenda for the next engagement meeting from them.
    Open - you decide1Documents, transcripts, playbooks
  2. Product-led B2B: Which customers renew within 60 days? For each, draft a note from the real history for approval.
  3. Agency: Give me an overview for our client Y: number of contacts, open deals and activities in the last 30 days.
Decided1
02
Getting started

How do we get started in practice?

  1. Inventory: which objects, fields and relationships does your team really use?
  2. First Model: map the standard objects and declare only what survives the inventory.
  3. Import: CSV for contacts, companies, deals and notes; the rest with support in the design-partner programme.
  4. Connect Claude: add the API token to the MCP configuration of Claude Desktop or Claude Code.
  5. Set up one Process with approval: renewal notes that a person approves, for example.
  6. Expand: every new Model brings new tools, without an integration project of its own.

Typical prompts with the Podma Connector

  1. Inventory: Walk me through the inventory: which object types and fields do we have, how many records, and which fields are mostly empty?
  2. First Model: Based on this inventory, propose a first, lean Model for our company, with a reason for every field.
  3. First win: Prepare me for my next customer meeting, from Podma data only, with sources.
03
Getting started

Which questions can I ask Claude once CMP42 is connected?

Goal Example prompt for Claude What happens behind the scenes
Clarify ownership “Who at Aspen owns the Q4 renewal?” Claude follows the edge contact owns deal and names the records it read
Pipeline overview “List all deals in the renewal stage and flag the ones at risk.” list_deals_at_stage, then get_context per deal
Meeting preparation “Prepare me for the call with company X: open deals, latest notes, contact people.” reads company, deals, notes and documents
Follow-up “Summarise the call and save the note on the company.” write_note, versioned and attributed to the agent
Prioritisation “Which three accounts are worth a call this week, and why?” reads the signal records; answers with date and source per signal
Resolve a contradiction “Which lifecycle stage applies to company Y?” both source records with origin and timestamp
Compliance “Which projects hold Art. 9 data, and under which DPA?” filter on the project record

Tool names are examples from the product description; the details may still change during the Alpha.

04
Getting started

What is the design-partner programme?

Twelve companies take part in the Alpha for six months, starting in October 2026. They get direct Slack access to the founder, a monthly co-design session, prioritised feature requests and the right to propose and name one MCP tool that ships to every future customer. In return protagx expects real use with real data and honest feedback; no testimonials, referrals or logos are asked for.

The price is a one-time EUR 6,900 net for the six months. From the seventh month, design partners can continue at the permanent founder rate of EUR 149 net per workspace per month, billed annually. Applying starts with a 15-minute call; applications close in December 2026.

Typical prompts with the Podma Connector

  1. Find your own MCP tool: Which query do I keep asking you in several steps? Turn it into a proposal for an MCP tool of our own, with name, inputs and result.
05
Getting started

What does the roadmap look like?

Period Milestone
October 2026 Alpha with the first design partners, weekly releases
Q4 2026 First case studies, stabilised MCP tool surface
Q1 2027 Beta with a waiting list, self-hosting image, HubSpot connector with OAuth
H1 2027 General availability (GA), published standard pricing
Q3 2027 SOC 2 Type II planned

protagx takes questions this FAQ does not answer at hello@protagx.com.

Decided1

¹ Planned for a future release.

// What comes next

WHAT THE NEXT RELEASES BRING.

The capabilities the answers above name as announced. Vote for the ones you want first — a click, no address, no sign-up. The order is fixed, never sorted by votes.

  • Open - you decide1Buying signals as objects
  • Open - you decide1Funnel metrics over MCP
  • Open - you decide1Approved forecast snapshots
  • Open - you decide1Follow-up sequences
  • Open - you decide1Event-driven processes
  • Open - you decide1Products and line items on deals
  • Open - you decide1Customer journey timeline
  • Open - you decide1Frozen audiences
  • Open - you decide1Documents, transcripts, playbooks
  • Open - you decide1Field-level access per MCP token
  • Open - you decide1Record history over MCP
  • Open - you decide1Reminders before retention ends
  • Open - you decide1Source records with provenance

¹ Planned for a future release.

Your requirement is not on the list

Then it is the one we want to hear. A design partner names what their week needs, and that is what moves to the front — the vote above only ranks what we have already decided. Suggest a time, or go straight to the Design Partner programme.