Why 70%+ of Companies Are Turning to Staff Augmentation for AI and DevSecOps Roles

Here is the problem those numbers describe. You have budget approved for an AI engineer and a DevSecOps specialist. You have had it for four months. The roles are still open.

As a Project Manager at Acquaint Softtech, I hear this version of the story most weeks, and the shape is always the same. The requisition goes out. Good candidates appear, then take a competing offer. The shortlist thins. Six weeks in, someone suggests widening the brief, and now you are interviewing generalists for a specialist role because the specialist market has priced you out.

---

Meanwhile the work does not wait. The AI feature slips a quarter. Your best senior engineer absorbs the security reviews on top of their own delivery, so both jobs get done at seventy percent. The compliance deadline does not move.

That is what the widely quoted 70% figure really describes: not a trend everyone is following, but a gap between what companies need and what the hiring market can supply. This article covers what the data actually says, why these two roles broke hiring, what an open seat costs you, and how to fill it.

What the 70% Figure Actually Means

Four independent studies measured different things in this market and landed within a few points of each other.

Source	Finding
Gartner IT workforce research	Over 70% of CIOs name skills shortages as their top barrier to transformation
Gartner, 2025	Close to 70% of generative AI projects stalled on talent gaps and complexity
ISC2 Workforce Study 2025 (16,029 practitioners)	69% engaged in AI adoption; skills, not headcount, is now the pressing concern
Skillsoft / Pluralsight, 2025	76% of employers report difficulty filling AI roles despite bigger budgets

Be careful how you quote this. None of these studies says “70% of companies use staff augmentation.” They measure a skills gap of roughly that size in these disciplines. Augmentation is the common response, not the number being counted. The honest version is also the stronger one in a budget meeting, because “everyone is doing it” invites the reply that everyone else is not us.

Why AI and DevSecOps Are the Hardest Roles to Fill

The US Bureau of Labor Statistics projects 29% employment growth for information security analysts between 2024 and 2034, against roughly 4% across all occupations. That is about seven times the national average, with around 17,300 openings a year. BLS attributes part of that growth directly to rising AI use, which is the link most hiring plans miss: your AI build and your security hire are one problem arriving from two directions.

Across computer and IT occupations generally, BLS projects roughly 317,700 openings each year. Supply is not keeping pace, and the people who hold these skills already have offers.

The security profession says the same thing about itself. The ISC2 Cybersecurity Workforce Study, covering 16,029 practitioners, found that the most pressing concern is no longer headcount but skills. Adding a body does not help if that body cannot do AI security. Their data puts AI and cloud security as the two most in-demand skill areas.

The scarcest profile of all

AI side	DevSecOps side
Machine learning and MLOps engineers	DevSecOps engineers embedding security into CI/CD
AI solution architects	Cloud security specialists
GenAI application developers	Compliance-fluent engineers (PCI DSS, GDPR)
Data engineers and data scientists	AI security specialists, an emerging category

Look at the bottom rows. The scarcest profile sits at the intersection: engineers who can build AI systems and secure them. That job title barely existed two years ago, so no established hiring pipeline produces it. 

This is exactly the gap our AI and ML engineers are placed into, usually alongside a security-fluent DevOps engineer rather than instead of one.

What an Open Specialist Role Costs You

Most teams compare an hourly rate against a salary. That is the wrong comparison, because it ignores what the empty seat costs while you search.

Line item	Permanent hire	Augmented engineer
Time to productive work	3 to 6 months	3 to 10 business days
Recruiting cost	$25,000 to $50,000	None
Fully loaded cost	Salary plus 25 to 40% overhead	Contracted rate, no overhead
Cost while the seat is empty	Carried for the whole search	Effectively zero
Cost if the fit is wrong	Months to spot, months to replace	Replaced within days
Cost when the peak passes	Permanent	Scales down with the engagement

The fourth row decides most cases. Work out what the blocked delivery costs you per month, then compare it to an augmented rate for the same period. That single calculation ends the internal debate faster than any argument about talent markets, because it turns an abstract shortage into a number your finance team already understands.

How Staff Augmentation Solves It

Augmentation means external engineers working inside your team, under your management, in your tools. You direct the work. It extends your capacity rather than handing a project away.

Solving the speed problem

A permanent senior AI hire takes about three months to close. An augmented engineer starts in days. That is the entire value proposition when a deadline is fixed, and it is why IT staff augmentation is the first move for most teams with a blocked roadmap. Our own onboarding runs within 48 hours of an agreed brief, with a one-week risk-free trial so you find out early whether the fit is real.

Solving the skills-mismatch problem

Hiring permanently assumes the required skill stays stable long enough to justify the commitment. In AI and security that is currently false. Bringing in engineers who have already shipped production AI in a regulated environment means you get judgement, not just capacity. 

When the need is specifically model and pipeline work, teams hire AI and ML engineers directly; when it is securing the CI/CD path around that work, they hire DevOps engineers with security fluency. Most real projects need both.

Solving the finite-work problem

Building an AI capability is heavy work for two quarters and light work afterwards. Hiring permanently for the peak leaves you paying through the trough. Where the peak lasts longer and covers a whole workstream, a dedicated software development team fits better than individual seats, and where the scope is already fixed and you would rather buy milestones than capacity, software development outsourcing delivers the same work against an agreed plan.

Solving the after-launch problem

The mistake I see most often is treating knowledge transfer as goodwill. Documentation, runbooks, and paired delivery should be written into the scope, or they are the first thing to slip when a deadline tightens. Once the build is live and your team owns it, support and maintenance keeps models tuned and security patched as threats and systems change.

When Augmentation Is the Wrong Answer

We sell this service, so treat this as the more useful half of the article.

1. The capability is permanently core to your product. If your AI models are the product, own that expertise. Augment to accelerate, but plan the permanent hire alongside it.

2. Nobody internal can review the work. A specialist with no internal counterpart produces work nobody can maintain. Assign someone to pair before you start.

3. The requirement is not defined yet. Augmentation is expensive discovery. Scoping first is far cheaper than paying senior engineers to work out what you want.

4. You need permanent 3am accountability. On-call ownership is a role, not a task. Augmentation covers build phases well and permanent operational ownership poorly.

If two or more apply, a short discovery workshop or an outside architecture opinion through virtual CTO services is the better first spend.

Case Study: Hybopay Finance

A Clutch-verified engagement that sits exactly at the AI and security intersection. Full review on our Clutch profile; comparable work in our case studies.

Client	Hybopay Finance Ltd., AI-driven lending, Dublin, Ireland
Reviewer	Gerhard Drobits, Chief Executive Officer
Engagement	AI development and custom software development, 6 to 10 engineers
Timeline	October 2025 to April 2026
Clutch rating	5.0 overall: Quality 5.0, Schedule 5.0, Cost 5.0, Willing to Refer 5.0

The problem

A small internal team was spending its time on manual document verification instead of training the AI agents that were the actual product. In automated lending, a minor decimal error compounds through risk calculations and reconciliation, so accuracy could not be relaxed to gain speed. They needed specialist AI engineers shipping code quickly, without a hiring cycle.

The solution

An augmented squad embedded directly into their GitHub and deployment workflows built a document ingestion pipeline: the AI reads a document, extracts the financial variables, converts them to structured data, and flags anything uncertain. Crucially, when confidence drops below threshold, the pipeline isolates that entry and surfaces it in a review console for a human, while everything else keeps moving.

The result

Documents that used to sit unverified for hours are now parsed and validated within moments. Data entry errors effectively disappeared, so credit agents work from clean data. In the client's own framing, this is what let a lean team operate at a scale their headcount alone never could. Every scheduled milestone was hit, including a mid-project schema change for multi-currency lending that was refactored over a single weekend.

What we would do differently

The client named it themselves. Our engineers first built the API endpoints using standard REST conventions, but their system runs on specialised schemas, so time went into refactoring the transport layer. An upfront integration specification during onboarding week would have avoided it. If you are augmenting into a system with non-standard conventions, write that document before the first line of code.

What It Costs in 2026

Indicative ranges for senior AI and DevSecOps talent, to help you budget rather than quote you.

Region	Senior specialist rate	Monthly equivalent
United States	$100 to $200 / hour	$16,000 to $32,000
United Kingdom	£80 to £150 / hour	£12,800 to £24,000
European Union	€85 to €160 / hour	€13,600 to €25,600
Australia	A$110 to A$200 / hour	A$17,600 to A$32,000
Acquaint (offshore)	$25 to $49 / hour	From $3,200 / month

Two caveats. AI engineers who can ship production systems commonly command 30 to 50% above general contractor rates wherever you source them. And offshore delivery at 50 to 70% below Western rates is only a saving if the seniority genuinely matches, so compare engineers rather than rates and insist on meeting the specific person first.

How to Run Your First Augmented Hire

Step	What to do	Why it matters
1. Define the outcome	Write what must be true in 90 days, not a job description	Specialists are hired against outcomes, not seats
2. Assign a counterpart	Name the internal engineer who will pair and inherit	Without this, knowledge leaves with the engagement
3. Prepare onboarding	Glossary, architecture notes, integration conventions	The most common first-fortnight friction
4. Contract the handover	Documentation and runbooks as deliverables	Otherwise it slips when deadlines tighten
5. Use the trial properly	Real work in week one, not reading	Fit is visible in five days if you let it be
6. Review at 30 days	Compare against the 90-day outcome from step one	Cheap to correct now, expensive later

Frequently Asked Questions