SD Infotech: Decide what AI should do in Talent Acquisition, and what people must

The problem

Companies are buying recruiting AI before defining what it should be allowed to decide.

Keyword screening rejects people who can do the job and passes people who only list the words. Adding AI on top of an unclear requisition automates the same mistakes, faster.

3 of 11

qualified candidates found by a keyword screen in our screening pilot. The evidence-based screen found 10.

candidates rejected by the evidence-based screen without a person looking. The keyword screen rejected 14.

40+

separate tasks and decisions in a typical recruiting process. Most AI decisions are made at the level of "screening" or "sourcing".

Pilot figures come from a controlled test on a constructed résumé set for one role. A live-requisition pilot is how we measure it in your organization.

The cost of getting it wrong

Before you invest more in recruiting AI, know what it's actually doing.

An AI tool can be technically impressive and still be the wrong investment for your hiring operation. Before committing more budget, expanding automation or presenting AI capabilities to enterprise buyers, establish what the technology actually does and whether the evidence supports its use.

What an EHDA review helps you gain

Less wasted AI investment

Find AI capabilities that duplicate what you already have, target poorly defined problems, or need process fixes before more money goes in. Employers get evidence to challenge vendor claims.

Real automation opportunities

See which repetitive work automation can safely take off recruiters' desks, and where a person's review has to stay.

Qualified candidates protected from unsupported rejection

Find where AI can screen out, rank down or filter candidates, and whether a person with the evidence checks it first.

More productive recruiting

Put recruiter time on the decisions that need judgment, with oversight and records that hold up when a decision is questioned.

For AI vendors, the same review gives your sales team documented, independent answers to enterprise buyers' due-diligence questions. A review identifies where value and risk sit; any savings depend on what you decide to do with the findings. We do not promise savings figures.

See the two-week pilots

The method

EHDA™: Evidence-Based Hiring Decision Architecture

Seven layers that take a hiring need from a vague request to a decision someone can defend. Layer 3 is the heart: what counts as proof that a person can do the work.

Layer 1

Hiring need clarity

What must this hire change, and by when?

Layer 2

Role and success architecture

Outcomes and capabilities, not inherited JD lines.

Layer 3 · the heart

Evidence architecture

Separates what a candidate has shown from what they only claim.

Layer 4

Decision architecture

Designs how the decision is made. It does not make it.

Layer 5

Human / AI allocation

Who can, should and must act on each task.

Layer 6

AI readiness and value

Scored per use case, never one company-wide score.

Layer 7

Validation and learning

Outcome evidence that earns AI more autonomy over time.

The Human-AI Opportunity Matrix

Every task is assessed against four questions. Autonomy is earned, never assumed.

Decision readinessIs the evidence behind this task clean and validated?

AI suitabilityCan AI do this reliably and explainably?

Decision riskWhat happens to a candidate if it is wrong?

Human accountabilityWho must answer for the outcome?

What the labels mean

AutomateAI acts. People audit samples.

HybridAI recommends or drafts. A recruiter approves before anything reaches a candidate.

HumanA person decides and is accountable. AI may organize the evidence.

Decisions that reject a candidate stay with an accountable person.

Two ways to start

One two-week pilot. Fixed scope. Built for employers or AI vendors.

The EHDA Decision Review (pilot) is the same method and the same standard, applied to your recruiting operation or to your product. Scope is agreed in writing at kickoff.

For employers · 2 weeks · fixed scope

EHDA Decision Review (pilot) of your recruiting operation

For TA, HR transformation and HR technology leaders deciding what their AI tools should be allowed to do.

One business unit and hiring workflow
Up to 2 job families
Up to 6 AI tools or AI-assisted recruiting tasks
Your systems as configured today
One consolidated round of factual corrections

Book a 30-minute discovery call

For AI vendors · 2 weeks · fixed scope

EHDA Decision Review (pilot) of your product

For recruiting AI vendors whose enterprise buyers ask what the product decides on its own. More for vendors

One product version
Up to 2 modules
Up to 6 AI features
One configuration
One consolidated round of factual corrections

Book a 30-minute discovery call

What you receive

A written assessment of each agreed AI feature or recruiting task, labelled Automate, Hybrid or Human, with oversight findings, record-keeping findings and specific recommendations.
A 60-minute findings review with your designated stakeholders.
A one-page summary: for your leadership if you are an employer, shareable with customers if you are a vendor.

What it is not

Independent findings and recommendations, not guaranteed cost savings, compliance certification, a bias audit or software implementation. Findings separate what was claimed, documented, observed and could not be verified.

After a pilot, and scoped by proposal: the full TA Automation Map diagnostic with a 90-day plan, a live-requisition screening pilot, and a readiness re-score every 6 months.

Who this is for

Mid-sized companies that hire every week and are weighing AI tools.

You run active hiring with a TA team and an ATS.
You are evaluating, piloting or already using AI recruiting tools.
You have no formal rule for what those tools may decide.
You need to explain your hiring decisions to leaders, candidates or regulators.

Who we work with

VP or Head of Talent Acquisition

Owns hiring quality and speed, and the decision to adopt AI.

HR transformation leader

Redesigning the TA operating model and needs a defensible allocation of work.

HR technology leader

Selecting and configuring recruiting AI and accountable for how it behaves.

For AI recruiting vendors

Your buyers are asking what your AI is allowed to decide. Give them an independent answer.

Enterprise TA and procurement teams now ask, before they sign: which decisions does the product make on its own, where does a recruiter approve, and who is accountable when a candidate is rejected? An EHDA Decision Review gives your sales team a clear, third-party answer.

Book a 30-minute discovery call

EHDA Decision Review (pilot) · 2 weeks

Each AI feature mapped to the hiring decision it influences, and labelled Automate, Hybrid or Human.
Where human oversight sits in your product today, and where buyers will expect it.
The evidence your product records behind each recommendation, and the gaps.
A written review for buyer due diligence, plus a one-page summary you can share with customers.

Fixed scope: one product version, up to two modules, up to six AI features, one configuration and one round of factual corrections.

Independent by design. Some features come back "keep this human", and that's what makes the review worth showing a buyer.

Proof

We test the method before we sell it.

Pilot 01: the same 22 résumés, 11 of them qualified, run through two screens.

Keyword screen3 of 11 found

14 candidates rejected by the machine, no person involved.

Evidence screen10 of 11 found

0 rejected by the machine. Every rejection left to a person.

Pilot 01 · screening

Keyword screen vs evidence screen

One role, 22 résumés, two screens. The keyword screen found 3 of 11 qualified candidates; the evidence screen surfaced 10, with every rejection left to a person. The evidence screen also missed: it advanced one candidate with invented metrics, which is why interviews still verify evidence.

A controlled demonstration on a constructed résumé set for one role, not a measure of results in a live hiring operation. Method and data available on request.

Field study · 10 real applications

Strong evidence, zero replies

Ten real job descriptions run against the résumés actually submitted. Half passed a literal keyword screen and four showed strong evidence for the role. None received a reply. Keyword matching explains only part of what filters candidates out.

Founder's own applications, analysed with consent.

Paps MuppaFounder, SD Infotech LLCTalent Acquisition SME since 2001

About

Built by a Talent Acquisition SME who has run the process from the inside.

Paps Muppa founded SD Infotech LLC to fix how companies decide who gets hired, and how much of that work AI should do. He has recruited since 2001: technical and executive search, leading recruiting teams, and in-house leadership hiring at JPMorgan Chase.

EHDA came from watching good candidates rejected by filters that never read their evidence, and from watching companies adopt AI without first defining what it was allowed to decide.

165hires at JPMorgan Chase, 2022–2024, including MDs, EDs and VPs

85%offer acceptance on those searches

42days average time to fill across 50+ open requisitions at peak

15recruiters led across distributed teams

Start here

Thirty minutes. One recruiting task you're unsure about. A straight answer on whether AI should touch it.

Tell us whether you are an employer weighing AI tools or a vendor answering buyer questions, and which tool or task worries you. We'll tell you straight whether a two-week pilot is worth it. Not ready to talk? Get a free readiness snapshot first.

Book a 30-minute discovery callpaps.muppa@sdinfo-tech.com