Beyond the Summary: Structured Data in Hiring

A summary is text. And text is good for humans. You read it, you understand it, you act on it. But for your CRM, text is useless.

Your CRM runs on fields. Numbers, dates, picklists, checkboxes, lookups. When you write a summary after a conversation and paste it in a notes field, there’s technically information in your CRM. But you can’t do anything with it. You can’t filter on “all candidates with a salary expectation between 50,000 and 70,000 euros.” You can’t report on availability. You can’t trigger an automatic workflow based on a skill.

The summary is the beginning. Structured data is where the value lives.

The difference between text and data

Let’s make it concrete. In a conversation, a candidate says: “I’m looking for something around 65,000 gross per year. I can start in six weeks. I have my PMP certification and speak fluent Dutch, English, and German.”

In a summary, this becomes: “Candidate is looking for a salary around 65,000 gross, is available in six weeks, has PMP certification and speaks three languages.”

That’s clear and useful for you as a recruiter. But for your CRM, it’s an unstructured string. To get value from it, you have to manually:

• Enter 65,000 in the salary field
• Calculate the date six weeks from now and enter it in the availability field
• Select PMP in the certifications picklist
• Add Dutch, English, German to the language fields

That takes time. And it’s error-prone. And it’s exactly the type of work recruiters don’t want to do.

How Simply extracts structured data

Simply’s data extraction goes beyond summaries. The AI system analyzes the conversation and automatically recognizes data points that can be written as structured fields to your CRM.

Back to our example. Simply hears the same sentence as you. But it does more than summarize:

• Salary expectation: 65,000 EUR gross/year (numeric, with currency and periodicity)
• Availability date: [current date + 6 weeks] (as a date object)
• Certifications: PMP (matched against your CRM taxonomy)
• Languages: Dutch (fluent), English (fluent), German (fluent) (as structured language fields)

Each value is written in the correct format to the correct CRM field. Not as text in a notes field, but as actual field values.

Smart recognition: context is everything

It sounds simple: “find numbers and put them in fields.” But it’s not. The difficulty is in the context.

A candidate says: “I currently earn 58 and want to get to 65.” What’s 58? What’s 65? Euros? Thousand euros? Per month? Per year? Simply understands the context. In a salary discussion in the Netherlands, where the topic is gross annual income, the AI interprets “58” as “58,000 EUR gross per year.”

Or a candidate says: “I was at Deloitte for two years and before that four years at KPMG.” Simply doesn’t just extract the company names but also the duration. And it places them in the correct order in your work experience fields.

Or, more subtly: a candidate says “I’m open to Amsterdam or Utrecht, but not Rotterdam.” Simply recognizes three locations and classifies them: two preferred, one excluded.

The validation system: trust with control

Not every extraction is equally certain. A clearly stated phone number (“My number is 06-12345678”) is more certain than an implied salary (“something in the neighborhood of 70”). That’s why Simply uses a validation system with color indicators:

• Green: high confidence. The AI is certain about the value and format. Automatically filled.
• Orange: medium confidence. The AI found a value but there’s some uncertainty. Please verify.

This system gives you the best of both worlds. High-confidence values save you time: there is nothing to review. The low scores keep you alert to exactly the points where things could go wrong.

And when in doubt: every field is clickable. You can listen back to the exact moment in the conversation where the AI sourced the value. That’s the transparency that makes Simply unique.

What data Simply extracts

The list is longer than you’d expect. Simply recognizes and extracts, among others:

Contact details

• Phone number (with automatic formatting to correct format)
• Email address
• LinkedIn profile
• Home address or region

Financial information

• Salary expectation (gross/net, per month/year, with currency)
• Current salary
• Hourly rate (for freelancers)
• Travel allowance or secondary benefits

Availability

• Start date
• Notice period
• Hours per week (full-time/part-time)
• Willingness to travel

Professional

• Skills (matched against your CRM taxonomy)
• Certifications
• Work experience (company, role, duration)
• Education level
• Languages and proficiency

Personal

• Hobbies and interests
• Family situation (if relevant and shared by the candidate)
• Personal motivation and values

Impact on your CRM data quality

The most underestimated problem in recruitment is CRM data quality. Most CRMs are a mess of empty fields, outdated information, and inconsistent notation. “65k” in one record, “65,000” in another, “approximately 65” in a third.

Simply solves this by formatting all extracted data according to your CRM’s rules. If your salary field is numeric, a number goes in. If your skills field is a picklist, only values from that picklist are used. No free text, no inconsistencies.

After six months of Simply, your CRM data is cleaner than it’s ever been. Not because you spend more time on it, but less.

From data to decisions

Structured data in your CRM opens possibilities that are impossible with unstructured notes:

• Matching: automatically search for candidates who meet a specific profile (salary, location, skills, availability)
• Reporting: what’s the average salary expectation in IT this quarter? What’s the average notice period for senior candidates?
• Automation: trigger workflows based on data fields (candidate available within two weeks? Automatically add to shortlist)

Combined with dashboards and reporting, you can see patterns across your entire database. Which skills are mentioned most often? Are salary expectations rising or falling? How many candidates are immediately available versus in three months?

Where the extracted data lands

Simply writes extracted data straight to the candidate record. No sync delay and no second system to keep in step, because Simply is the system of record: the extraction and the record are the same product. Field mapping is configured by you: you decide which Simply output goes to which field, including the fields you added to the model yourself.

What connects outward are the tools around the conversation — Gmail, Outlook, Google Calendar, Zoom, Slack and HubSpot — so mail, agenda and meetings hang off that same record. And if you have specific needs not covered by default, Simply offers API customization options.

Structured data as a foundation for predictive recruitment

When you consistently collect structured data across hundreds of conversations, predictive capabilities emerge. You can analyze which combination of skills, experience, and salary expectations most frequently leads to a successful placement with a specific type of client. Those patterns are invisible in unstructured summaries but become clear once the data is standardized.

A concrete example: suppose your structured data shows that candidates with more than five years of supply chain management experience and a salary expectation between 65,000 and 75,000 euros have a placement rate of 82% with logistics companies in the Randstad area. That information enables you to immediately identify the most promising candidates for a similar new vacancy, even before the first conversation takes place.

This transforms recruitment from a reactive to a proactive process. Instead of waiting for the right candidate, you can predict which candidates in your database have the highest success rate for each new assignment. Structured data is the fuel that makes this possible.

Maintaining data quality in structured extraction

Structured data is only valuable when quality is consistent. Simply uses a validation system with green and orange indicators to show the reliability of each extracted value. Green means the system has high confidence in the data’s correctness. Orange means human verification is recommended.

This validation system prevents inaccurate data from automatically entering your CRM. A salary expectation the candidate mentioned hesitantly or where ambiguity existed gets flagged for review. An availability date that was explicitly confirmed goes through directly. This way you maintain the speed of automation without sacrificing accuracy.