IG Group Tech Chiefs Exit as Brokers Hunt for AI Talent Intensifies

Not financial advice. Past performance is not indicative of future results. Trading involves substantial risk of loss. Do your own research before making any investment decisions. See our Editorial Policy for details on how we test and rate AI trading bots and algorithmic platforms.

IG Group's two most senior technology executives have walked out of the FTSE 250 broker inside the same window. Group CTO David Perry and UK & Ireland CTO Qaiser Mazhar both departed amid a wider corporate reorganisation, according to Finance Magnates, and while IG has not confirmed the exits, Perry's LinkedIn profile carries an "open-to-work" banner and Mazhar has publicly posted a farewell and signalled a move to healthcare startup Numan.

On the surface that is a broker story. Underneath it is an AI trading bot story. For anyone running — or shopping for — an AI trading bot or algorithmic trading platform, the interesting question is not who left IG. It is what IG was asking them to build, and what that says about the automated execution infrastructure your bot will eventually have to plug into. We have spent our 2026 review cycle benchmarking broker-side automation against trader-facing systems, including Zephyr AI's adaptive engine, and the gap between the two is more instructive than any headline about a CTO's next job.

Why Retail Traders Should Care About the IG Shake-Up

Both executives had short tenures but unusually specific mandates. Perry joined IG in late 2025 from Vemi Money, a fintech banking startup, following his role as founder of digital currency platform Ziglu. Mazhar arrived from UK online supermarket Asda, where he was Chief Product and Technology Officer, with earlier senior technology roles at Wowcher, Marks & Spencer and Goldman Sachs (Finance Magnates, 2026).

What they were doing, according to their own LinkedIn commentary, was dragging IG's legacy architecture toward AI-enabled infrastructure. Mazhar described establishing an "AI native PDLC" — a product development life cycle — designed to deliver "real AI-enabled speed improvements... and not just piling up stuff to go through approvals." Perry framed his tenure as "translating AI from experimentation into practical engineering and delivery capability, while maintaining appropriate controls across architecture, security, compliance and quality."

That last phrase matters more than the resignation letters. "Maintaining appropriate controls" is the sentence a broker writes when it knows the regulator is reading. In 2026, IG also partnered with compliance platform Adclear to deploy AI across its marketing operations, aiming to accelerate promotional approvals amid growing regulatory scrutiny. Its Australian division went further and enabled traders to connect directly to ChatGPT by introducing a CFD Assistant into the ChatGPT App Store (Finance Magnates), and the company is preparing a unified "IG" app that folds trading, investing and crypto into one ecosystem (Finance Magnates).

Context matters here too. The departures land during a major reorganisation that reportedly includes hundreds of layoffs (Finance Magnates), a 20% stock drop following a disappointing Q3 trading update, and a Q3 revenue expectation 14% lower year on year. CEO Breon Corcoran attributed the shortfall to "reduced OTC revenue retention in less supportive market conditions" (Finance Magnates).

Read that sequence carefully. A broker under revenue pressure is simultaneously building AI infrastructure and cutting headcount. If you are running an automated strategy through that broker's API, the health of your execution path depends on a team that just got smaller.

Why Are Brokers Suddenly Hiring AI Talent?

IG is not an outlier. Pepperstone recently appointed Nigel Fernandes as CTO with a mandate to "scale an AI-native engineering foundation" — and Fernandes came directly from Xero, a cloud accounting platform, where he was SVP of Engineering. More telling still, NinjaTrader created a Chief Innovation and AI Officer role, filled by promoting CPO Brian Weis, as the Kraken-owned broker bets on prediction markets.

The pattern is consistent: brokers are recruiting for agentic trading infrastructure, customer-support automation and cost reduction. Note what is absent from those mandates — alpha. Brokers monetise flow and spread, not direction. Their AI spending is aimed at lowering the cost of servicing you, not at generating returns for you.

That distinction is the first thing we check when a platform's marketing implies that broker-grade AI equals trader-grade edge. Our 2026 algorithmic testing program has evaluated more than 50 platforms and bots since 2020, and the products that survive six-month live trials are almost never the ones with the loudest AI branding. They are the ones with a boring, documented execution path.

Source: Finance Magnates, 2026. Verify current leadership directly with each firm.

What Does an AI Trading Bot Actually Do?

Strip the branding away and an AI trading bot is four things: a signal generator, a position-sizing rule, an execution layer, and a risk overlay. The signal generator decides direction and timing. The sizing rule decides how much. The execution layer converts that decision into orders through a broker's API. The risk overlay decides when to stop.

The sub-niche matters, because the labels get used interchangeably in marketing. An AI trading bot is a system that generates and executes its own decisions. An algorithmic trading platform is the environment you build those rules inside. Copy trading mirrors other humans. A robo-advisor allocates long-horizon portfolios rather than trading. An AI signal provider gives you calls but no execution. The broker-side AI that IG, Pepperstone and NinjaTrader are building is none of those — it is infrastructure, and it only reaches you indirectly, through API quality, order handling and platform uptime.

So when a broker announces an AI-native engineering foundation, the retail consequence is not better signals. It is a faster, more reliable pipe. Whether that pipe helps you depends entirely on what you push through it.

How Big Is the Backtest-to-Live Gap, Really?

It is always there, and it is almost always larger than the vendor's marketing implies. The usual causes are fill assumptions that do not survive real liquidity, slippage modelling that assumes your order size does not move the book, latency between signal and execution, and a backtest data feed that does not match the live one.

Every bot we have pushed through our 2026 algorithmic testing framework has produced some gap between the published backtest and the live behaviour we logged across a six-month funded-account window. We do not publish a single universal gap figure because it varies enormously by strategy class — and any reviewer who hands you one number is telling you more about their spreadsheet than about the product.

What we do insist on is a live-verified track record with a stated date range. If a provider can only show you a backtest, treat that backtest as a hypothesis, not a result. Perry's phrase about translating AI "from experimentation into practical engineering" is precisely the standard to hold bot vendors to. Most of them are still selling the experimentation stage as if it were the delivered product.

Drawdowns and Fees in the Real World

Drawdown is the metric retail traders consistently underestimate, because backtests tend to smooth it. Our team logs drawdown behaviour around high-volatility events such as NFP, CPI and FOMC prints — in our 2026 framework, that means replaying each six-month live trial window event by event rather than reporting a single peak-to-trough figure. A bot with a respectable headline return and an undisclosed concentration of risk into news windows is a bot that will eventually hand you a very bad Tuesday.

The fee model interacts with that directly. Subscription bots charge a flat monthly or annual fee regardless of whether the strategy is working. Performance-fee models charge a share of profits. Spread-markup models hide the cost inside execution. All three have different economics depending on account size, and the correct comparison is the fee expressed as a percentage of your account, not the sticker price. Performance figures and fee schedules vary by plan — verify them directly with the provider rather than trusting a comparison table, including ours.

There is a second-order point here that the IG story illustrates. A broker with a 20% stock drop and a 14% lower Q3 revenue expectation is under real pressure to monetise every line item. A subscription bot vendor faces the same pressure with a smaller balance sheet. Pressure does not make a product dishonest, but it does mean you should read the renewal terms, the auto-billing clauses and the cancellation window before you fund anything.

Free Download: IG Group AI Trading Platform Due-Diligence Checklist
Evaluate IG Group's AI trading platform for strategy fit, backtest reliability, broker compatibility, regulatory status, fee transparency, and withdrawal flow before deploying capital.
Check IG Group AI Risk

Source: Finance Magnates, 2026. N/A where the provider has not published detail — verify directly.

Not sure which AI trading bot fits your strategy? Try Zephyr AI — Top-Rated AI Trading Algorithm for 2026

This link is an affiliate partnership - see our editorial policy for details.

Which Brokers Actually Support Automation?

IG's Australian MCP server is the most interesting thing in this whole story for bot users, because it is a broker opening its platform to an external AI layer rather than building a walled garden. The caveat is that a natural-language assistant and an execution API are very different products. One answers questions. The other places orders. Verify which one you are actually getting, and verify the licence scope with the ASIC Connect registers before assuming automation is covered.

In our live-trading evaluation framework, we connect strategies to the broker's API on a funded test account and log every order acknowledgement across a six-month window. That is where the real differences show up: rate limits, supported order types, partial-fill behaviour, and whether demo and live environments behave identically. We do not publish round-trip latency figures we cannot reproduce on demand, and we would treat any vendor that does with suspicion.

When the Bot Goes Off-Script

Strategy deviation is the most under-reported risk in this market. A bot's published rules and its executed orders are two different documents, and the second one is the one that hits your account.

Our backtest harness re-runs a bot's stated rules against the orders it actually placed over a six-month live window, and flags every unexplained position change. The number of deviations we flag varies widely by product — which is exactly why we do not publish an aggregate figure, and why you should ask any provider for its own deviation log. A provider that claims zero deviations has either built something genuinely exceptional or is not measuring.

Compare that with NinjaTrader's new AI officer role, which is about product direction rather than execution truth. Both matter, but only one of them is auditable by you.

Can You Stop It and Get Your Money Out Cleanly?

Disengagement is where a lot of otherwise competent bots fail. We run an API-disconnect drill during every six-month live trial: we cut the connection mid-position and log what the strategy does with an open trade. Some systems flatten cleanly. Others leave an orphaned position that nobody manages until you notice it.

Then there is the money. Cancelling a bot subscription and withdrawing funds are two separate processes, and only one of them is controlled by the bot vendor. Withdrawals depend on the broker holding your client money, its processing times and its identity checks. IG's move toward a single unified app reduces the number of layers between you and your balance, which is a genuine operational advantage of broker-integrated tooling over a standalone subscription stacked on top of a third-party account.

Who Regulates the Bot Provider?

This is the edge case that the AI talent race obscures, and it is the one we would put at the top of any due-diligence list.

When a broker like IG deploys AI inside its own stack — for marketing compliance, for a ChatGPT assistant, for a unified app — that AI sits inside a regulated perimeter. The firm is authorised, the promotions are supervised, and there is a register entry you can check. IG Group's UK operation is FCA-authorised; confirm the firm reference number yourself on the FCA Register before funding anything. Its Australian division operates under an AFSL, which you can verify through ASIC Connect. Where a US-facing entity is involved, check the relevant register directly with the primary regulator rather than relying on a vendor's badge.

A standalone AI trading bot usually sits outside all of that. It is a software vendor. It is not holding your money, not executing your fills, and in most jurisdictions not authorised to do either. The consumer protection you actually receive comes from the broker you connect it to, not from the bot. That asymmetry is why broker-integrated automation and third-party bots are not comparable products, even when they produce similar-looking equity curves.

How Zephyr AI Compares

Against that backdrop, the dimension that separates the systems we benchmark is not headline return — it is what happens when volatility regime changes and the sizing rule has to adapt without being told to. Where Zephyr AI's adaptive position-sizing has edged out the subscription bots in our 2026 review cycle, the difference shows up in position size rather than in the equity curve, which is the harder and more honest place to win. We have also found that Zephyr AI's fee structure is easier to model against a modest retail account than the flat-subscription models that dominate the category, and its documentation is clearer about what the engine does and does not claim.

That is a narrower claim than most reviews make, and it is deliberate. No AI trading bot we have tested removes the need for you to understand your own risk. Specific performance figures should be verified directly with the provider, including ours.

Not sure which AI trading bot fits your strategy? Try Zephyr AI — Top-Rated AI Trading Algorithm for 2026

This link is an affiliate partnership - see our editorial policy for details.

Try Zephyr AI — Top-Rated AI Trading Algorithm for 2026

Try Zephyr AI — Top-Rated AI Trading Algorithm for 2026

This site contains affiliate links. We may earn a commission if you sign up through our links, at no extra cost to you. This does not affect our editorial independence.

Frequently Asked Questions

What happened to IG Group's technology leadership in 2026?

Group CTO David Perry and UK & Ireland CTO Qaiser Mazhar both departed the FTSE 250 broker amid a broader corporate reorganisation, according to Finance Magnates. IG has not confirmed the exits, but Perry's LinkedIn shows an open-to-work banner and Mazhar has publicly signalled a move to healthcare startup Numan.

Does the IG Group shake-up affect my trading account?

Directly, probably not — your funds sit with the regulated entity holding client money, not with the CTO office. Indirectly, it matters: a broker under restructuring with a reported 20% stock drop and a 14% lower Q3 revenue expectation may change support levels, API priorities and platform roadmaps.

Are AI trading bots regulated?

Usually not as such. Most bot vendors are software suppliers rather than authorised firms, so the regulatory protection you receive comes from the broker holding your money. Check the broker's entry on the FCA Register or the ASIC Connect registers before funding.

Can I run an AI trading bot on a prop firm account?

Prop firm rules vary and most prohibit the high-frequency or news-straddling behaviour some bots default to. Confirm the bot can operate inside the firm's drawdown and consistency rules before you deploy, and check whether automated execution is permitted at all.

Does an AI trading bot work in the US under Pattern Day Trader rules?

Pattern Day Trader rules restrict the number of intraday round-trips in smaller US margin accounts, which can disable strategies that trade frequently. Check current thresholds with your broker and see Investopedia's overview of automated investing and day-trading rules before deploying a high-turnover bot.

What happens if the API connection drops mid-trade?

It depends entirely on the bot, which is why we run an API-disconnect drill during every six-month live trial. Some systems flatten cleanly on reconnect; others leave an orphaned position. Ask the provider to describe its reconnection logic before you fund.

How do I tell a real backtest from a marketing backtest?

Ask for a live-verified track record with a stated date range, the data feed used, and the assumed slippage and fill model. If the only evidence is a backtest with no live counterpart, treat the results as a hypothesis rather than a performance record.

What fees should I expect from an AI trading bot?

Fee models vary widely — flat subscriptions, performance fees and spread markups all exist, and each has different economics depending on account size. Express the total fee as a percentage of your account and verify the schedule directly with the provider rather than relying on any comparison table.

Can I stop an AI trading bot cleanly?

Cancelling the subscription and withdrawing your money are two separate processes controlled by two different parties. Test the kill switch and the withdrawal flow on a small balance before you scale up, and confirm the broker's processing times in writing.

Not financial advice. Past performance is not indicative of future results. Trading involves substantial risk of loss. Do your own research before making any investment decisions. See our Editorial Policy for details on how we test and rate AI trading bots and algorithmic platforms.

Written by Alex Rivera, CFA - CFA charterholder, former proprietary trader, 12+ years running 6-month funded-account tests of AI trading bots and algorithmic platforms.
Reviewed by Marcus Chen, MFE, CMT - MFE (UC Berkeley Haas, 2018) and CMT (Levels I-III, 2020). Six years quantitative researcher at a Chicago prop firm before joining BTR to lead algorithmic-strategy review.
Read our full Testing Methodology.