AI readiness for scale-ups.

What is AI readiness for a scale-up?

AI readiness is a company’s ability to turn AI into measurable revenue or cost outcomes, safely and repeatedly. It is not the number of tools in use or pilots under way. A ready scale-up knows which use cases matter, has the data and engineering to run them in production, can explain them to customers and has a named person accountable for the risk.

The gap between use and value is wide. McKinsey’s State of AI 2025 survey found that 88% of respondents report regular AI use in at least one business function, yet nearly two-thirds say their organisations have not begun scaling AI across the enterprise, and 39% attribute any EBIT impact to AI. BCG’s The Widening AI Value Gap (September 2025) put 5% of companies in its group generating value at scale, with 60% reporting hardly any material value. Adoption is also uneven: Eurostat reports that 19.95% of EU enterprises with ten or more employees used AI technologies in 2025, against 55.03% of large enterprises.

The assessment below is written for a leadership team to work through together. Answer each question with evidence, not opinion, and place each area on the maturity scale.

How do you score AI maturity?

Use four stages and score each of the six areas separately. Most scale-ups sit at different stages in different areas, and the weakest area usually sets the pace.

• Exploring. Individuals use AI tools on their own initiative. There is no agreed list of use cases, no budget line and no policy. Value is anecdotal.
• Piloting. Two or three use cases are being tested with a defined owner and success measure. Data is pulled by hand. Costs and risks are understood loosely.
• Operating. At least one AI capability runs in production with monitoring, an evaluation routine and a known cost per transaction. A usage policy exists and staff are trained.
• Scaling. AI is part of how the company builds, sells and runs. Use cases are added through a repeatable process, governance is routine and the board tracks AI outcomes alongside other operating metrics.

Moving one stage in an area within a quarter is a realistic goal; jumping from exploring to scaling tends to produce expensive pilots.

Which AI use cases should a scale-up pursue first?

Pick the use cases where AI changes revenue or cost in a way you can measure, not the ones that demo well. Strategy is the first area because every other area depends on it.

• Revenue or cost. For each candidate use case, which metric moves (conversion, expansion revenue, gross margin, support cost per ticket, time to ship), and by how much would it need to move to matter?
• Customer pull. Have customers asked for it, or would they pay more, renew or expand because of it?
• Advantage. Does the use case depend on something the company has that competitors do not, such as proprietary data or a workflow position?
• Focus. Is there a short list of no more than three priority use cases, each with a named owner?
• Stopping rule. What result after a set period ends a pilot?

Ready looks like: a written short list of use cases tied to operating metrics, owners and stopping rules, agreed by the leadership team and reported to the board. Shaping that list is strategy work of the kind covered by 5X Strategic Mentorship & Global Network.

Is your data ready for AI?

Data is ready when the company owns it, can reach it, trusts its quality and has the right to use it for AI. Gartner said in February 2025 that, through 2026, organisations will abandon 60% of AI projects that are not supported by AI-ready data (Gartner).

• Ownership. Which data does the company own outright, and which sits with customers, partners or vendors under contract terms that limit its use?
• Consent. Do customer contracts and privacy notices allow the data to be used to train, fine-tune or ground AI features? Is personal data handled under the GDPR or the relevant local law?
• Quality. Is there a known owner and a quality check for the data each priority use case depends on?
• Access. Can an engineer reach the data through a governed pipeline, or does each use case start with manual exports?
• Feedback. Does the product capture how users respond to AI output, so the system can be evaluated and improved?

Ready looks like: a data inventory for the priority use cases, with ownership, consent position, quality owner and access route recorded for each.

How should engineering and architecture support AI?

The architecture should let the company change models, measure quality and control running costs without rebuilding the product. That is an engineering discipline, and it maps directly to 2X Software Engineering & Tech Excellence.

• Build or buy. For each use case, which parts are bought (models, platforms) and which are built (data pipelines, integration, user experience)? Is the reasoning written down?
• Model choice. Has the team compared more than one model on its own tasks, including smaller and open-weight options where they fit?
• Evaluation. Is there a test set and an automated evaluation that runs before every change reaches customers?
• Inference cost. What does each AI transaction cost to run, and how does that cost change gross margin at ten times current usage?
• Vendor lock-in. Could the company switch model provider within weeks if prices, terms or quality changed?

Ready looks like: a model abstraction layer, an evaluation suite, cost per transaction on a dashboard and a documented build-or-buy decision for each use case. Teams using AI coding agents should do so with review, testing and security controls; see AI coding agents in engineering teams.

A caution on agents in particular: Gartner predicted in June 2025 that over 40% of agentic AI projects will be cancelled by the end of 2027, citing escalating costs, unclear business value or inadequate risk controls (Gartner). The cost and evaluation questions above address exactly those reasons.

How does AI change go-to-market?

AI changes what customers buy, how they find vendors and how the company should price. The go-to-market questions sit with 1X Revenue Engineering, where revenue is the outcome measured.

• Product story. Can the sales team explain in one sentence what the AI does for the customer, in the customer’s terms, without overclaiming?
• Pricing. Is AI value captured in pricing (a tier, usage, or outcomes), and does the price cover the inference cost at scale? See how AI is changing software pricing.
• AI discovery. When a buyer asks an AI assistant about the company’s category, does the company appear, and is the description accurate?
• Sales and marketing use. Is AI in demand generation measured against qualified pipeline?
• Buyer questions. Can the company answer customer security and AI governance questionnaires quickly and consistently?

Ready looks like: AI features described accurately in the product story, a pricing model that protects margin and a regular check of how AI assistants describe the company.

Are your people ready to work with AI?

People are ready when staff have the skills to use AI well, managers redesign work around it and there is a plan for the change. McKinsey’s 2025 survey found that its AI high performers were nearly three times as likely as others to fundamentally redesign workflows when deploying AI, which makes ways of working, not tools, the deciding factor.

• AI literacy. Do staff who use or oversee AI understand what it can and cannot do? Under the EU AI Act, the AI literacy obligation for providers and deployers has applied since 2 February 2025, and the Commission notes that no certificate is required (European Commission AI literacy Q&A).
• Skills. Which roles need deeper skills (engineering, data, product), and is that capability being hired, trained or partnered for?
• Workflow redesign. For each priority use case, has someone redrawn the process, rather than adding AI to the old one?
• Change management. Do teams know how roles will change, and is there a channel to raise concerns?

Ready looks like: a training record, a usage policy staff have read, and a redesigned workflow for each use case in production. Peer experience from founders who have been through the change is part of what 5X Strategic Mentorship & Global Network provides.

What AI risk and governance does a board need?

The board needs a named accountable owner, a current view of regulatory exposure and a short AI risk register reviewed on a regular cycle. This is governance work that maps to 4X Governance & Back Office.

• Accountable owner. Which executive is accountable for AI outcomes and AI risk, and how often do they report to the board?
• EU AI Act exposure. Which AI features fall into which risk category, and is the company a provider, a deployer or both? See what the EU AI Act means for scale-ups. Founders should take legal advice on their own position.
• Security. Are prompt injection, data leakage through AI tools and access controls for AI systems covered in the security programme?
• Intellectual property. Are the terms for model outputs, training data and code generated by AI understood and reflected in customer contracts?
• Human oversight. For each AI decision that affects customers, where does a person review, override or stop the system?
• Incident handling. Is there a process for AI errors that reach customers, including how they are recorded and disclosed?

Ready looks like: an AI policy approved by the board, a risk register with owners, a mapped view of EU AI Act exposure and AI as a standing board agenda item.

What should the first 90 days look like?

A 90-day plan moves the company one stage forward in its weakest areas and puts one use case into production, starting from exploring or piloting.

Days 1 to 30: assess and choose

1. Run this assessment with the leadership team and score each area on the four-stage scale.
2. Name the accountable executive and agree how AI will be reported to the board.
3. Agree a short list of no more than three use cases, each tied to a revenue or cost metric, with an owner and a stopping rule.
4. Publish an interim AI usage policy and start AI literacy training for staff who use AI.

Days 31 to 60: build the foundations

1. Complete the data inventory for the chosen use cases, including consent and access.
2. Set up an evaluation suite and measure cost per transaction for the first use case.
3. Map AI features against EU AI Act risk categories and start the AI risk register.
4. Redesign the workflow around the first use case with the team that will run it.

Days 61 to 90: ship and review

1. Put the first use case into production with monitoring and human oversight in place.
2. Update pricing and the product story if the use case is customer-facing.
3. Report results against the agreed metric to the board, then decide to scale, change or stop.
4. Re-score the six areas and set the next 90-day plan.

For a wider view of operating readiness beyond AI, the scale-up readiness assessment covers the commercial, engineering and governance foundations that AI work depends on.

How does TGC Capital Partners work with founders on AI readiness?

TGC Capital Partners is Your Growth Engineering Partner, the strategic investment arm of Gateway Group of Companies, a software engineering company since 1997 with more than 2,000 people and offices in 15 countries. TGC invests in software and technology founders across Europe, North America and India, and works with their teams through the 6X Factors where each one fits. For AI readiness that is usually engineering (2X), revenue (1X) or governance (4X). Revenue is the outcome we measure, not a promise.

Frequently asked questions

What is an AI readiness assessment?

An AI readiness assessment is a structured set of questions a founder and board answer to judge whether the company can turn AI into revenue or cost outcomes, not demos. It covers six areas: strategy and use-case selection, data, engineering and architecture, go-to-market, people and ways of working, and risk and governance. Each area is placed on a maturity scale from exploring to scaling.

How long does an AI readiness assessment take?

A first pass takes a leadership team one or two working sessions if the right people are in the room: the founder, the technical leader, the commercial leader and whoever owns finance and legal matters. The value is in the evidence behind each answer, so expect follow-up work to check data access, inference costs and contracts before the board discusses the result.

Who should own AI in a scale-up?

One named executive should be accountable for AI outcomes and AI risk, reporting to the board on both. In most scale-ups this is the CEO or the technical leader, supported by a product owner for each use case. Shared ownership without a single accountable person tends to produce pilots that never reach production.

Does the EU AI Act apply to a scale-up outside the EU?

It can. The EU AI Act applies to providers that place AI systems on the EU market and to deployers using AI in the EU, wherever the company is based. Founders in North America or India with European customers should map their AI features against the Act’s risk categories and take legal advice on their position.

Should a scale-up build its own AI models or buy them?

Most scale-ups start by buying access to foundation models and building the parts that carry their advantage: proprietary data, workflow integration, evaluation and the user experience. Training or fine-tuning a model makes sense when the company has distinctive data and a measured quality gap that off-the-shelf models cannot close. The decision should be revisited as model prices and capabilities change.

What does AI-ready look like at board level?

The board sees a short list of AI use cases tied to revenue or cost metrics, the cost of running each one, the risk register for AI features, and a named owner. It asks for evidence, such as evaluation results and customer adoption, rather than demos. AI is a standing item on the board agenda, not a one-off presentation.

Speak with a partner 2X Software Engineering & Tech Excellence

Related reading: The EU AI Act for scale-ups · AI coding agents in engineering teams · AI and outcome-based software pricing · Scale-up readiness assessment · The 6X Factors