Signals Flow: The GTM System I Built to Take CREflow to Market

Building CREflow—a CRM and deal-flow platform for commercial real estate—meant I had to solve two problems at the same time:

1. Build a product that helps commercial real estate professionals manage data, workflows, and opportunities more effectively.
2. Build a repeatable, measurable way to identify the right market, reach the right people, and learn from every outbound interaction.

The second problem became Signals Flow.

Signals Flow started as the internal GTM operating system I needed to take CREflow to market. It turns a fragmented outbound stack into a structured workflow for defining an ideal customer profile, finding and evaluating accounts, identifying the right buyers, creating controlled personalization, launching campaigns, and feeding campaign outcomes back into the system.

I did not build it because there were no tools available. I built it because the tools were useful individually but did not give me a reliable system of record for the decisions happening between data sourcing and outbound execution.

The problem: outbound was fragmented

The initial CREflow outbound motion used a familiar modern GTM stack:

• LinkedIn Sales Navigator for market and persona research
• Apollo and Clay for account/contact data and enrichment
• n8n for workflow automation and integrations
• Smartlead for outbound campaign execution
• LLM APIs for research synthesis, data cleanup, copy support, and personalization
• Supabase and webhooks for application data, workflow state, and event-driven updates

That stack could produce lists and launch campaigns. But as activity increased, the operating model became harder to manage.

The real outbound decision-making lived across spreadsheets, Clay tables, formulas, enrichment credits, prompt outputs, tabs, and handoffs between tools. An account could be enriched in one place, scored in another, researched in a third, and pushed to a sequencer with limited visibility into why it qualified or what happened after it entered a campaign.

For CREflow, that was a problem because commercial real estate is not a simple volume-outbound market. A contact may look relevant based on title, company size, or LinkedIn activity but still be the wrong person, market, property type, or operating model for the product. A managing partner, acquisitions lead, broker-owner, developer, asset manager, or CRE operator can each have materially different workflows and reasons to care about a CRM or deal-flow system.

I needed a way to make that judgment process visible, repeatable, and improvable.

The goal was not “send more cold email.”

The goal was to build a controlled process that answers:

• Is this company within CREflow’s target customer profile?
• Why does it fit—or not fit?
• Which people at the company actually influence the relevant workflow?
• What evidence supports the outreach angle?
• Has the contact and email been verified before campaign entry?
• What personalization is helpful versus generic AI-generated fluff?
• What happened after the lead entered an outbound sequence?
• How should positive, negative, or non-response outcomes change future targeting?

That is the gap Signals Flow fills.

Why I built Signals Flow

I wanted a dedicated outbound control plane: a workspace where the logic and quality standards of GTM live independently from the tools used to source data or send email.

Clay, Apollo, LinkedIn Sales Navigator, and similar systems remain valuable inputs. Signals Flow is not designed to replace every tool in a GTM stack. Instead, it owns the decision layer after lead intake.

The core design principle is simple:

Data providers supply information. Sequencers send messages. Signals Flow determines what deserves to move between them.

Rather than treating an account or person as a row in a temporary spreadsheet, Signals Flow treats each prospect as a structured lead object with a clear history:

• Source and import context
• Company profile and ICP attributes
• Enrichment outputs
• AI research notes
• Fit score and qualification rationale
• Relevant contacts and buyer roles
• Personalization variables
• Verification status
• Campaign and sequence status
• Engagement and reply outcomes
• Manual reviewer feedback

This makes the workflow more auditable. Instead of asking, “Why was this lead sent to Smartlead?” I can see the evidence, scoring outcome, qualification status, research context, and approval logic that moved it forward.

That approach reflects the GTM systems I have been building end-to-end: capture, enrichment, scoring, routing, outreach, CRM or database synchronization, and measurement—not just list building or prompt-generated copy.

How Signals Flow works

Signals Flow follows a gated workflow:

ICP → Enrich → Score → Verify → Personalize → Push → Engage

Each stage exists to prevent bad-fit or under-researched accounts from advancing simply because data was available.

1. Define the CREflow ICP

The first step is converting CREflow’s market strategy into usable operating rules.

For a commercial-real-estate SaaS product, an ICP cannot be only a title filter such as “commercial real estate professional.” The market needs to be segmented based on practical criteria such as:

• Firm type: brokerage, investment firm, developer, owner-operator, asset manager, lender, or property-management organization
• Market focus: commercial, industrial, multifamily, office, retail, land, or mixed-use
• Geographic footprint and target markets
• Team size and business complexity
• Deal volume or portfolio workflow signals
• Current operational maturity
• Evidence of fragmented deal tracking, manual data entry, disconnected systems, or an active growth motion
• Likely decision-maker, champion, or operational user

Signals Flow can use GTM strategy documents as an input. An LLM-assisted extraction workflow turns those documents into structured ICP fields, scoring conditions, messaging guidance, and tone rules. Those rules remain editable; AI can accelerate the translation of strategy into operations, but it should not be treated as the final authority on who belongs in the market.

For CREflow, this matters because the product is built around commercial-real-estate workflows rather than generic CRM use cases. The system needs to distinguish between a company that simply has a real-estate title somewhere on LinkedIn and a firm that has a meaningful deal-flow, relationship, property, or pipeline-management need.

2. Import and enrich target accounts

Target accounts can enter the workflow through existing outbound tools and research sources, including Clay, Apollo, and manually assembled prospect lists.

Signals Flow then enriches the company and its relevant people using public web and LinkedIn-oriented research sources. The point is not to collect every available data point. It is to assemble the evidence needed to decide whether the account deserves outreach.

The enrichment layer is designed to surface information such as:

• What the company does
• Its relevant commercial-real-estate specialization
• Markets served
• Firm size and likely operating complexity
• Recent activity, hiring, growth, transactions, or expansion signals where available
• Company website positioning and service model
• Decision-makers and operational stakeholders
• Contact details and verification status
• Potential reasons CREflow could be relevant now

This replaces the manual process of opening dozens of browser tabs, copying notes into a spreadsheet, and hoping the information stays tied to the right contact.

3. Score fit with evidence

Once an account is enriched, Signals Flow evaluates fit.

The scoring process uses AI-assisted research, but it is designed to produce a reasoned decision—not a black-box score. Each account is evaluated against the operating ICP and returned as:

• Fit
• Not fit
• Needs review

The system generates supporting research notes explaining why an account was qualified or rejected. For example, a CRE firm may be a strong candidate because it has a distributed acquisitions team, visible market expansion, an active property pipeline, and signs of managing opportunities across several stakeholders. Another company may be rejected because it is outside the target vertical, too small for the current motion, not commercially focused, or has no apparent need for a deal-flow platform.

This stage reduces a common outbound failure mode: treating technically valid contact data as proof of commercial relevance.

Data validity and product fit are separate questions. Signals Flow is designed to make both explicit.

4. Keep humans in the loop

One of the most important decisions I made was to preserve manual feedback and review gates.

AI can research, normalize, summarize, classify, and identify patterns at a scale that is impractical to do manually. But it can still miss the nuance that matters in a specific vertical, geography, or buyer motion.

For CREflow, that nuance is important. A brokerage with ten agents, a regional investment firm, and a sophisticated owner-operator may all appear relevant on the surface, but their workflows, buying urgency, and product requirements can be very different.

Signals Flow therefore allows manual Fit / Not-fit labels to be added to a subset of accounts. Those labels become a feedback mechanism for improving future AI scoring and refining the actual ICP.

The intended operating model is human-in-the-loop GTM:

• AI handles repetitive research and first-pass qualification.
• The operator reviews edge cases and strategically important accounts.
• Manual feedback improves the quality of future scoring.
• Only accounts that meet the required quality threshold advance to outreach.

This is not a system built to remove judgment. It is built to apply judgment more consistently and reserve human time for decisions that actually need it.

Controlled personalization, not AI spam

Personalization is often where outbound systems fail.

At one extreme, teams send generic templates at high volume. At the other, they allow an LLM to generate unrestricted copy with inconsistent claims, weak relevance, and variable brand voice. Neither approach is ideal—especially for a product like CREflow, where credibility and vertical relevance matter.

Signals Flow creates structured lead-intelligence variables before copy generation. These variables can include the firm’s market focus, team structure, website positioning, recent signals, relevant operations, and an identified workflow hypothesis.

The copy layer then works within strict templates and rules:

• Defined messaging objectives
• Approved tone and brand voice
• Clear claims that can be supported
• Specific personalization fields
• Required or prohibited language
• Output formats suitable for campaign sequencing
• Final coherence and quality checks before handoff

The purpose is to make outreach more relevant without creating a free-form AI system that writes something different—and potentially off-brand—for every prospect.

For CREflow, this makes it possible to test targeted messages around practical problems: disorganized deal tracking, fragmented property and contact data, weak pipeline visibility, inconsistent follow-up, or difficulties coordinating opportunities across a commercial-real-estate team.

The message should be tailored to the prospect’s likely operating reality, but it should remain grounded, concise, and credible.

Closing the loop with Smartlead

Once a lead has passed the qualification and review gates, Signals Flow hands it off to Smartlead for campaign execution.

The important part is that the workflow does not stop at the sequencer.

Campaign events and reply webhooks flow back into Signals Flow so I can connect the outbound outcome to the data and reasoning that created the outreach in the first place. That enables a more useful feedback loop:

1. A target account enters from Apollo, Clay, Sales Navigator, or a curated list.
2. The company and relevant contacts are enriched.
3. Signals Flow scores fit and records why.
4. A reviewer can confirm, reject, or correct the result.
5. Lead intelligence generates approved personalization variables.
6. The lead enters a Smartlead sequence.
7. Delivery, engagement, and reply events return to the workspace.
8. Positive replies, objections, disqualifications, and no-response patterns inform future segmentation, scoring, and messaging.

This is important for CREflow because early-stage GTM is fundamentally a learning process. The objective is not simply to maximize emails sent. It is to learn which segments respond, which pain points resonate, which buyer titles convert, and where the product positioning needs refinement.

How I use it for CREflow

Signals Flow is the GTM research, qualification, and outbound layer for CREflow.

CREflow itself is a commercial-real-estate CRM and deal-flow product built around CRE-specific workflows. Its technical foundation includes AI-enabled processing of public-record data, property-evaluation pitch generation, a Supabase real-time data layer, lead enrichment pathways, and event-driven communications.

Signals Flow supports CREflow’s path to market by creating a disciplined account-based outbound motion around the people and firms most likely to benefit from the product.

In practice, I use it to:

• Turn CREflow’s target-market hypotheses into explicit ICP and scoring rules.
• Build focused account lists instead of broad, title-based prospect lists.
• Research firms and identify likely decision-makers, champions, and users.
• Validate that the company has a credible commercial-real-estate workflow fit before outreach.
• Generate relevant research context for outreach without manually researching every account from scratch.
• Keep messaging consistent while still tailoring it to the firm, segment, and likely workflow pain point.
• Send only qualified and verified leads into Smartlead campaigns.
• Monitor replies and campaign outcomes in relation to account quality, segment, and message.
• Use real market feedback to refine the product narrative, ICP, and roadmap.

For example, if a particular group of firms repeatedly responds to messaging about fragmented deal tracking and relationship management, that is not just an outreach insight. It is product-market feedback. It can inform CREflow’s positioning, landing pages, demo narrative, feature prioritization, and future segmentation.

Likewise, if a segment has high email engagement but low positive-reply quality, the problem may not be deliverability. It may indicate that the ICP, contact selection, timing, or value proposition needs adjustment.

Signals Flow creates the operational visibility to diagnose that difference.

What I learned building it

The biggest lesson was that a connected GTM stack is not necessarily an operating system.

It is relatively easy to wire APIs together, move a lead from one platform to another, or ask an LLM to write an email. It is much harder to create a dependable process that preserves context, controls quality, supports human review, tracks cost, and learns from outcomes.

A real outbound system needs more than automations. It needs:

• A clear system of record for account and lead decisions
• Explicit definitions of fit and disqualification
• Structured research evidence
• Validation before campaign entry
• Controlled personalization rules
• Event-driven campaign feedback
• A way for operators to correct the machine
• Visibility into what is working, what is wasting enrichment spend, and what needs to change

That is what Signals Flow is becoming.

Where Signals Flow is going

Signals Flow began as an internal system for bringing CREflow to market. I am now evolving it toward a multi-tenant platform for teams that run active outbound motions and want greater control over quality, research, personalization, and feedback loops.

The near-term roadmap includes:

• Multi-provider enrichment waterfalls
• Lookalike account sourcing
• Missing-email discovery and verification
• Expanded intent and signal detection beyond LinkedIn activity
• Native Clay and Apollo synchronization
• More robust feedback loops between manual qualification, AI scoring, and campaign outcomes
• Continued improvement of structured lead objects, reporting, and outbound quality controls

Because enrichment, API, and LLM usage create real operating costs, access is intentionally invite-only while I validate the workflows with a small number of teams.

I am most interested in working with operators who already have an outbound motion in place and care about better qualification, cleaner GTM data, more useful personalization, and stronger feedback loops—not simply sending higher volumes of email.

The bottom line

I built Signals Flow because bringing CREflow to market required more than lead lists and sequencers.

It required an accountable workflow that converts market strategy into operational rules, turns raw account data into researched and verified opportunities, applies AI where it genuinely saves time, retains human judgment where it matters, and feeds actual campaign outcomes back into the next GTM decision.

For CREflow, Signals Flow is how I operationalize the path from “this may be the right market” to “this is a verified, relevant account with a clear reason to contact it now.”