System for finding and qualifying new contacts
From search to CRM, with quality checks and human review in one workflow.
A system that starts searches from an operations chat, gathers businesses from a mapping service, checks quality and duplicates, brings in a person when needed and records only ready contacts in the CRM.
- Role
- Workflow architecture · Qualification rules · Integrations · Verification
- Scope
- Search and candidate collection · Quality, duplicates and review · Lead creation and summary
- Where it works
- Operations chat · Mapping service and PostgreSQL · Internal CRM
Request, results and source were managed at separate moments.
The chat command connects the search to candidates gathered in PostgreSQL.
01
The project
The project connects commercial research with contact management. An operator describes the businesses to find in an operations chat; the system gathers candidates, prepares them for evaluation and records the ready ones in the CRM.
The work covered the search command, mapping-service collection, PostgreSQL preparation area, quality rules, duplicate matching, human review and Lead creation in the CRM.
02
The starting need
The need was to make searches producing many candidates with different levels of completeness and relevance genuinely usable. Every result needed to preserve its source, available contact details and reasons for its assessment.
Consistent quality and duplicate checks were needed before recording, while keeping human review for cases requiring judgement.
03
Challenges and solutions
Connect a commercial search request to the candidates found and their source.
- What was delivered
- An operations-chat command starts the mapping-service search and records candidates and the session in PostgreSQL.
- Outcome achieved
- Every candidate remains linked to the search that produced it and can be assessed within the same workflow.
Recognise the same business even when its details arrive in different forms.
- What was delivered
- The system compares place identifier, domain and phone number before preparing the Lead.
- Outcome achieved
- Existing candidates are caught before recording, keeping the contact base cleaner.
Use quality rules without replacing human judgement in cases needing assessment.
- What was delivered
- Score, relevance and completeness distinguish ready candidates from those sent for human review.
- Outcome achieved
- People focus on cases that genuinely need a decision while keeping the reasons for the assessment visible.
Bring a checked contact into the CRM and keep the operational summary connected.
- What was delivered
- The workflow creates the Lead in the CRM and sends the outcome to the operations chat against the same search session.
- Outcome achieved
- Search, assessment, recording and ownership become consecutive, traceable steps.
04
How it works
The operations-chat command starts a mapping-service search and stores candidates in PostgreSQL. They are normalised and matched by place identifier, domain and phone number.
Quality rules separate ready candidates from those needing review. After assessment, the system creates a Lead in the CRM and returns a chat summary linked to the search.
Key decisions
- 01
Quality before recording
- Why
- The CRM should receive candidates that are already compared and understandable.
- What changed
- Source, completeness and duplicates are assessed before Lead creation.
- 02
Focused human review
- Why
- Some candidates require context and judgement.
- What changed
- Rules prepare the case and people decide with information already organised.
Project journey
From search to qualified Lead
- Operations-chat command
Search described by the operator
- Mapping-service search
Businesses matching the request
- Candidates in PostgreSQL
Source and contact details organised
- Quality and duplicates
Score, place, domain and phone
- RouteRequirements completeReady candidate
Requirements already complete
- RouteAssessment requiredHuman review
Decision with organised context
- Lead in the CRM
Qualified contact recorded
- Operations-chat summary
Outcome linked to the search
Technical delivery details
What was delivered
- Chat commands to start research
- Mapping-service collection and PostgreSQL area
- Quality rules and duplicate matching
- Human candidate review
- CRM Lead creation and chat summary
05
Before and after
Commercial research
BeforeRequest, results and source were managed at separate moments.
AfterThe chat command connects the search to candidates gathered in PostgreSQL.
Related evidence282 files checked
Quality and duplicates
BeforeEach candidate required repeated checks before recording.
AfterRules, place identifier, domain and phone prepare a consistent assessment.
Related evidence282 files checked · Qualification and recording verified
Ownership
BeforeAssessment, CRM and operational summary were separate steps.
AfterThe Lead and chat summary remain linked to the same search session.
Related evidence282 files checked · Qualification and recording verified
06
Results achieved
Result delivered
Candidates qualified before entering the CRM
Source, quality, relevance and duplicates are assessed before Lead creation.
Related evidence282 files checked · Qualification and recording verified
Search and ownership in one workflow
Chat command, PostgreSQL candidates, review, CRM Lead and operational summary stay connected.
Related evidence282 files checked
How it was verified
Essential evidence
- 282 files checked — The local validator checked system configuration, workflows, schema and references.
- Qualification and recording verified — Rules connect quality checks, duplicate matching, review and Lead creation.
07
More work
If you recognise a similar challenge, we can start from your real context.
Discuss your project