Startup proposal deck

AI first interview app

A validation-first product concept for automating first-stage interviews, saving recruiters three hours per candidate, and only building after two companies commit.

First stage

Frontend Engineer

18:42 recorded
Candidate answer evidence
ID match passed

Government ID shown before interview. Recruiter can review the proof frame.

AI scorecard

Role fit86%
Communication74%
Evidence quality81%
Recommendation: invite to human interview

01 / Recruiting bottleneck

First-stage screening is still too manual.

Recruiters spend valuable time preparing questions, scheduling, repeating first-round interviews, writing summaries, and sending rejection emails.

AI adoption is now normal in business, but HR leaders are being pushed to show measurable efficiency instead of experimentation. Gartner says HR operations are under pressure to automate service delivery and upgrade productivity.1

3h

Target time saved from every first-stage interview.

20m

Candidate completes the AI interview anytime, anywhere.

2+

Paid pilots required before implementation starts.

02 / Market signal

The timing is right, but trust is the product.

88%

Organizations using AI in at least one business function, according to McKinsey.2

45%

US workers who report using AI at work, while SHRM flags a widening HR skills gap.3

AI agents

HR buyers are moving from chatbots to workflow agents that schedule, summarize, and follow up.1

Audit

Employment AI is regulated. Bias audits, notice, human review, and documentation are part of the sale.4

03 / Product wedge

One link replaces the repetitive first interview.

JD-to-question generation
Anytime interview link
20 minute structured screen
AI summary and scoring
Identity check before recording
Human review with video evidence
Calendar scheduling
Auto sorry email for failed candidates

Core promise

The recruiter reviews evidence, not raw scheduling chaos.

04 / Workflow

The interview agent handles the first stage end to end.

01

Prepare

Upload a JD, target level, must-have skills, and interview rubric.

02

Invite

Candidate receives a 20 minute first-stage interview link and calendar options.

03

Verify

Candidate shows an ID card and confirms live presence before recording.

04

Interview

AI asks structured questions and role-specific follow-ups by video, phone, or chat.

05

Score

Recruiter receives summary, evidence clips, scorecard, and recommended next step.

06

Act

Pass candidates move to human scheduling. Rejections get a polite automatic email.

05 / System design

A lean architecture that can start manual and become software.

Candidate experience

Interview link
ID capture
Video / phone / chat
Consent and reminders

Interview agent

JD parser
Question planner
Follow-up policy
Rubric-based scoring

Review workspace

Video evidence
Summary
Scorecard
Pass / reject / schedule

Trust layer

Audit log
Bias checks
Retention policy
Human override
Candidate portal
Agent orchestration API
Recruiter review app

Storage: encrypted interview video, transcript, scorecard, ID proof frame, and audit log.

Integrations: Google Calendar, Outlook, email, ATS export, and later phone or messaging channels.

MVP path: automate the candidate link and review report first; keep calibration human-assisted.

06 / Agentic roadmap

Start with video. Expand into phone, email, and messaging.

Asynchronous video

Best for screening evidence, communication, and recruiter review.

AI phone interview

Best for high-volume roles, low friction, and faster candidate response.

Email follow-up

Best for invites, reminders, rejection notes, and hiring-manager nudges.

Messaging agent

Best for candidate Q&A, rescheduling, and status updates.

07 / Competitive gap

In Mongolia, the real competitors are job boards and recruiter services, not AI-native interview products.

Zangia.mn

Mongolia's largest local hiring platform and employer marketplace.6

Where they are strong

Strongest local distribution and employer mindshare for finding applicants.

Gap we can use

Built for attracting candidates, not for automating structured first-round interviews or producing evidence-based interview reports.

Worki.mn

Mobile-friendly local jobs platform with company pages, direct applications, and AI-assisted salary tools.7

Where they are strong

Modern candidate UX and visible adoption by major Mongolian employers.

Gap we can use

Application flow is the main product. Interview workflow, scoring, and recruiter-side automation are not the wedge.

Mongolia Talent Network

Recruiter-led hiring, employer branding, PEO/EOR, and premium talent services for Mongolian employers.8

Where they are strong

High-trust local service model and strong credibility with established companies.

Gap we can use

Service-heavy and human-led. Less scalable and less standardized than AI-run first-round screening.

Mongolia Talent Network on Worki

Local employer profile showing Ulaanbaatar presence and long-running recruitment operations.9

Where they are strong

Confirms deep local market presence and employer-side hiring relationships.

Gap we can use

Still a recruiter-service model, not a software-native interview layer.

Zangia ecosystem tools

Adjacent local substitutes such as Testly and training products inside the Zangia ecosystem.10

Where they are strong

Easy add-on for employers already using a familiar Mongolian HR brand.

Gap we can use

Assessments and training do not replace a full conversational interview screen with scheduling, evidence clips, and human review.

Positioning takeaway

The wedge is not "more HR software." The wedge is a Mongolia-ready interview layer that sits after sourcing, automates first-round screening, and gives hiring managers evidence instead of extra admin.

08 / Compliance by design

No black box hiring decisions. Human review stays in control.

Consent and retention

Candidates see what is collected, why, and when video or ID evidence is deleted.

Human override

AI recommends. The hiring team makes the decision and can inspect every evidence clip.

Bias audit posture

Structured rubrics, score distribution checks, and documentation for regulated buyers.

No facial emotion scoring

ID verification is for authenticity only. Avoid pseudoscientific personality claims.

EEOC guidance and local AEDT laws make transparency, job-relatedness, and bias controls a sales requirement, not a feature afterthought.5

09 / Pricing

Price per interview first. Move to subscription after proof.

Pilot

10,000 MNT / interview

Pay as you go, first 10 interviews, white-glove setup

Starter

390,000 MNT / month

50 interviews, JD templates, calendar automation, email follow-ups

Growth

1,200,000 MNT / month

180 interviews, branded portal, reviewer seats, analytics export

Agency

Custom + usage

Multiple clients, shared pools, ATS export, dedicated onboarding

Validation offer

First 10 interviews as a paid pilot with manual white-glove setup.

10 / Go to market

Sell the promise before building the platform.

01

Pitch deck

Use this page as the core story for HR managers, founders, and recruiting agencies.

02

Landing page

Create one focused page with pricing, workflow, trust posture, and a booking link.

03

5 emails per day

Target HR teams in tech, finance, education, customer support, and agencies.

04

Calendly booking

Every email and deck CTA routes to a short pilot discovery call.

05

Manual MVP

Use AI tools plus human QA behind the scenes until two companies commit.

06

Buyer proof

Track replies, call bookings, pilot conversion, objections, and time-saved estimates.

11 / Decision rule

Do not build until two companies say yes.

Success condition: two paid pilot commitments.

A commitment can be a signed pilot, prepaid interview bundle, or written buyer agreement to run a first batch after delivery.

Measure objections

Trust, candidate experience, pricing, and integration concerns.

Measure willingness to pay

Validate 10,000 MNT per interview against volume discounts.

Measure urgency

Find teams with active hiring pipelines now, not someday.

Appendix

Research notes

Market-size numbers for AI hiring vary widely by definition. The stronger investor story is not a giant top-down number. It is a narrow wedge: recruiters already feel the pain, AI is accepted in workflows, and compliance makes audited human-in-the-loop products more credible.