Technology professional indemnity, cyber insurance, and directors and officers cover - the core components of technology insurance for AI software companies - are all written on a specialist basis. Underwriters assess each application on its merits, considering the nature of the technology, the clients it serves, the contractual obligations involved, and the security and governance practices of the business.

There is no standard rate per employee or per million pounds of revenue. Two AI companies with identical revenues may pay materially different premiums because their risk profiles - what they build, who they sell to, how they manage claims exposure - are fundamentally different.

When it comes to technology insurance costs for AI risks, underwriters are grappling with genuinely new questions: how do you price liability for autonomous decision-making? What does a model hallucination claim look like? How does regulatory exposure - including the EU AI Act where the business operates in EU jurisdiction - affect pricing? The pricing landscape is evolving, and specialist placement matters.

Technology professional indemnity - also called tech errors and omissions (Tech E&O) - is the core cover for AI software companies. The premium is influenced by:

• Revenue. Premium scales broadly with revenue because higher-revenue businesses have larger potential client exposures. A company with lower ARR and a company with significantly higher ARR face different indemnity exposures, and pricing reflects that.
• Nature of the AI system. Underwriters look closely at risk level. A system that makes autonomous decisions in a high-risk domain - healthcare triage, credit scoring, employment selection, critical infrastructure - attracts materially higher underwriting scrutiny than a B2B SaaS analytics tool with human oversight. For companies operating in EU jurisdiction, the EU AI Act's risk classifications inform this assessment; UK-only work is governed by a UK version of AI principles and sector-specific rules.
• Complexity of client contracts. Uncapped liability clauses, unlimited indemnity obligations, or negligence standards that go beyond reasonable care all drive premium higher. Contracts with appropriate liability caps represent a lower-risk position.
• Indemnity limit required. The higher the limit of indemnity you carry - whether because clients contractually require it or because your own risk assessment demands it - the higher the premium.
• Claims history. A company with prior PI claims will pay more than a claims-free business. The nature of any prior claims - their cause, resolution, and what the business did in response - will be examined closely.
• Whether the policy explicitly extends to AI liability. Some specialist Tech E&O products now include explicit extensions covering AI-generated outputs, algorithmic decision-making liability, and AI system failures. These extensions represent broader coverage, and premiums reflect that. If you are comparing standard tech PI quotes against specialist AI-extended policies, you are not comparing like for like - the specialist policy may cost more and provide substantially more relevant protection.

For a full breakdown of what technology PI covers and how to evaluate policy wording for AI companies, see our guide to professional indemnity insurance for software companies and AI developers.

Cyber insurance for AI companies is underwritten on the basis of specific data, systems, and security posture. Key factors include:

• Data volumes and sensitivity. Businesses that hold or process large volumes of personal data - particularly special category data such as health, financial, or biometric data - represent greater regulatory and financial exposure in a breach. AI training pipelines that aggregate personal data from multiple sources attract heightened scrutiny.
• Security certifications. ISO 27001 and Cyber Essentials (or Cyber Essentials Plus) are the two most commonly recognised certifications in UK cyber insurance underwriting. Holding both is a meaningful differentiator with underwriters and can materially improve both pricing and terms.
• Cyber security controls. Good cyber hygiene - multi-factor authentication (MFA), endpoint protection, privileged access management, and regular patching - is positively viewed by underwriters and can help reduce premiums.
• Revenue. Business interruption cover is sized relative to revenue. Higher revenue means a larger potential interruption loss.
• Third-party integration count. The more external APIs, cloud services, and data processors your systems connect to, the greater the supply chain risk. Underwriters assess the complexity of your integration landscape.
• Incident history. Prior cyber incidents - ransomware, data breaches, extortion demands - are material to underwriting. A clean history is a real pricing advantage.

For more detail on what cyber insurance covers, see our guide to cyber insurance for AI companies in the UK.

Directors and officers insurance premiums for AI startups are driven by:

• Company stage. Early-stage companies typically represent smaller absolute exposures. As a company raises significant capital, takes on institutional investors, and grows its director population, the exposure increases.
• Number of directors. More directors means more potential sources of claims. The composition of the board - executive directors, non-executives, advisers with formal roles - all factor into the assessment.
• Investor structure. Institutional investors bring formal governance expectations and, in some cases, greater willingness to pursue claims against directors. A VC-backed Series B company faces a different D&O exposure than a bootstrapped startup.
• Regulated sector exposure. If your AI company operates in financial services, healthcare, or another regulated sector, D&O premiums will reflect the additional regulatory investigation exposure this creates.

For a complete guide to D&O for AI founders, see: Directors and Officers Insurance for AI Startup Founders.

The limit of indemnity you choose has a direct relationship with your premium: higher limits cost more. But the relationship is not linear - moving from £1m to £2m does not double the premium.

For technology professional indemnity, the most common limits are:

• £1m - standard for early-stage companies without enterprise contracts specifying higher limits
• £2m - increasingly common as a minimum for mid-market enterprise and public sector contracts
• £5m - required by larger enterprise clients, regulated sector contracts, and businesses where a single incident could generate a significant claim

Enterprise contracts often specify a minimum indemnity limit. If a client requires £2m PI cover and you hold £1m, you may not be able to take on that contract without adjusting your cover. This is worth factoring into renewal conversations before contract negotiations, not after.

Underwriters reward businesses that demonstrate genuine risk management. Steps that improve your position:

Achieve ISO 27001. The single most impactful certification for technology insurance underwriting. It demonstrates systematic management of information security risk and is recognised across both cyber and PI underwriting.

Obtain Cyber Essentials Plus. The government-backed scheme provides verified assurance of basic security controls. Many cyber insurers ask about it directly.

Implement strong cyber controls. Multi-factor authentication (MFA), endpoint protection, privileged access management, and regular patching are looked on favourably by underwriters and can help reduce premiums for cyber and PI cover.

Document your AI governance framework. Underwriters want to see that AI governance exists: how models are tested and monitored, how bias and accuracy risks are assessed, how deployment decisions are made. A written policy is evidence of managed risk.

Review contracts proactively. Unlimited liability clauses and warranties you cannot practically fulfil are underwriting red flags. A legal review of standard client contracts with a focus on liability positions will pay for itself in insurance terms.

For AI companies building on generative models, additional cost factors apply. See: Insurance for Generative AI Products.

The technology insurance cost question cannot be answered with a single number - it depends entirely on your specific business profile. What can be done is to ensure your insurance is placed with underwriters who understand AI risk, that your risk profile is presented accurately, and that you are not paying for cover that does not respond to the risks you actually face.

Get a technology insurance quote from Taurus Risk - specialists who understand your risk.

The right premium for an AI company is the one that buys cover which responds when a claim is made. Working with specialist underwriters, presenting your risk profile clearly, and matching limits to contractual obligations are what move pricing in the right direction.