Task 2.3: Position AI for competitive advantage.

• AFrom being first to adopt each new foundation model as soon as it is released

  Incorrect

  Competitors can adopt the same models on the same day, so early access to a public model is a short-lived edge at best.

• BFrom negotiating a lower per-token price than competitors pay

  Incorrect

  Lower unit costs help margins, but they do not make the insurer's AI capabilities different from its competitors'.

• CFrom keeping all AI use cases secret until each is fully deployed

  Incorrect

  Secrecy can delay imitation briefly, but it does not create a capability that competitors cannot reproduce with the same models.

• DFrom its proprietary claims data applied to its own processes

  Correct

  AWS's guidance calls data the strategic differentiator for generative AI: models thrive on high-quality, context-rich, well-governed data, which competitors cannot obtain by subscribing to the same model.

When everyone can rent the same models, the difference lies in what you bring to them: proprietary data, processes and knowledge that competitors do not have. Investing in that data, its quality and its governance builds an advantage that is hard to copy.

• ACopy the largest competitor's announced AI projects, because they set the industry direction

  Incorrect

  Copying a competitor's announcements ignores the distributor's own strengths, data and constraints, and arrives second with the same capability.

• BCommission a custom model now so that the company is not left behind

  Incorrect

  Building before understanding the landscape and the opportunities risks investing in a capability that does not create an edge.

• CWait for competitors to publish proven results

  Incorrect

  Waiting for proof from others delays learning, and competitors rarely publish the details that would make their results reusable.

• DDocument the industry's AI applications and peers' lessons learned

  Correct

  AWS's maturity guidance starts with documented awareness of industry AI applications and opportunities, including learning from other companies' adoption experience at industry forums.

Positioning AI for advantage starts with an informed view of the landscape: what the industry is doing, what peers have learned and where the company's own strengths could make a difference. Priorities set from that view are more defensible than imitation or haste.

• APut most of the budget into generative design, where no rival has a lead yet and the upside is largest

  Incorrect

  Concentrating on an unproven area stakes the budget on uncertain returns and cannot deliver the savings the board wants within 12 months, while rivals keep their cost advantage.

• BFund the proven uses only, and start generative design once rivals have published their results

  Incorrect

  Waiting for others to prove the emerging area leaves the company without the experience it needs; with a two-year product cycle, it would fall behind if generative design pays off.

• CScale the proven uses to close the cost gap, and ring-fence a smaller budget for design experiments

  Correct

  AWS advises balancing a portfolio across short- and long-term outcomes and across proven, low-risk and experimental, higher-risk opportunities, within the organization's financial and schedule constraints.

• DSpend the first year on a company-wide AI platform, then choose the use cases to build on it

  Incorrect

  A platform year with no use case in production produces no savings in the 12-month window and delays time to value, which AWS encourages organizations to shorten.

Investment should follow how mature each use of AI is in the industry. Where AI is already common, the company has to close a gap quickly with proven uses; where it is still emerging, bounded experiments buy knowledge and options without staking the budget. A balanced portfolio meets the near-term target and keeps the long-term option open.

• AUse AI to summarize internal policy documents so staff find answers faster

  Incorrect

  Summarizing internal documents improves productivity. It is an efficiency gain inside the current business model.

• BAutomate the reconciliation of bank files in the back office

  Incorrect

  Back-office automation lowers operating cost but leaves how the company earns revenue unchanged.

• CPrice client contracts on outcomes delivered by AI agents

  Correct

  Outcome-based pricing ties payments to measurable business results. AWS describes the shift from fixed-cost models to outcome-based models as a transformation of the business model that AI agents make possible.

• DDeploy an assistant that drafts responses for the support team

  Incorrect

  Drafting support replies raises agent productivity, an operational improvement rather than a new business model.

• ESell a workforce-insights product built on anonymized payroll data

  Correct

  A new data-enabled offering is a new value proposition and revenue model, which is how AWS describes using technology to reach new customers and create new revenue.

AI can optimize the current business or change it. Transformation means new value propositions, new revenue streams or new ways of charging, such as pricing on outcomes or selling data-enabled products, rather than doing today's work more cheaply.

• AInsurer B's workflow integration, which keeps adapting and improving with use

  Correct

  AWS finds that successful organizations favor learning-capable systems over static tools and workflow integration over standalone tools, and that the systems that succeed retain context, adjust to changing business requirements and improve through operational experience. Months of such gains are something a later rival has to earn over time.

• BInsurer A's six-month head start in putting an assistant in front of claims staff

  Incorrect

  The lead is fixed in time: the tool has not improved since launch, so any rival on the same model can deploy an equivalent one. AWS links static tools that cannot adapt to high failure rates.

• CInsurer A's long system prompt, which spells out its claims rules in detail

  Incorrect

  A prompt is an input that guides the same model; it does not change the model or learn from use, and a rival can write equivalent instructions.

• DInsurer A's lower per-request prices from its capacity commitment with the provider

  Incorrect

  Lower prices for committed capacity or batch processing are published options open to every customer of the model, so rivals can obtain the same terms.

A model anyone can rent is not an advantage on its own, and neither are timing, instructions or prices that a rival can match. An assistant woven into the workflow that keeps context and adapts as the business changes gets better with every month of operation, and that accumulated improvement is what competitors cannot copy quickly.

• AWaiting is better, because early work will be wasted when rules change

  Incorrect

  Waiting leaves the startup scrambling when rules arrive, while prepared competitors move faster.

• BStaying ahead of compliance gives an edge once regulations take hold

  Correct

  AWS's guidance says organizations that stay ahead of compliance and regulatory frameworks with a security-first approach will have a competitive advantage once regulations take hold.

• CGovernance has no competitive effect, only a cost

  Incorrect

  Hospitals buying AI care about compliance; being ready earlier is a selling point.

• DThe startup should avoid regulated customers entirely

  Incorrect

  Avoiding regulated markets gives up its core customers.

In regulated markets, readiness is part of the product. Building governance and security ahead of final rules lets a company sell confidently as requirements arrive.

• AFund a price cut that matches the rivals with savings from AI automation

  Incorrect

  Savings from automation would only pay for the across-the-board price match, which would wipe out the margin; the move gives up margin on every subscriber, including the many who would have stayed.

• BUse AI-generated campaigns to win new subscribers in nearby regions

  Incorrect

  Reaching new customers is a growth move in AWS's strategy guidance, but it does nothing to stop the losses from the existing base, which is the CEO's goal for this year.

• CAdd an AI assistant that answers support questions around the clock

  Incorrect

  AWS ties good customer service to retention, but the exit surveys show that price, not service, is the main reason these subscribers leave, so the assistant does not address the cause.

• DPredict which subscribers may leave and target offers at them

  Correct

  AWS's Cloud Adoption Framework points to customer behavior insights for subscriber retention and to data's analytical value in predicting outcomes; offers aimed at likely leavers protect revenue without cutting every subscriber's price.

In a price war, a blanket price cut gives away margin on every customer, including those who would have stayed. Predictions built from the company's own customer behavior data let it spend retention offers where they change the outcome, turning data it already holds into measurable protection of revenue.

• AWait until a rival proves agents, then copy the proven design

  Incorrect

  AWS lists competitive advantage among the business drivers for agentic AI because early adopters gain faster insights, better service and adaptive operations; waiting hands that lead to whichever rival moves first.

• BInvest now, running agents as production services tied to outcomes

  Correct

  Early adopters gain strategic advantage, and AWS stresses that agents must run as production-grade services that deliver real business outcomes, such as the faster responses shippers choose carriers on.

• CFund many small agent experiments with no path to production

  Incorrect

  AWS warns that intelligence without integration yields novelty, not impact; experiments that never reach production deliver none of the service gains shippers notice.

• DReplace the whole dispatch team with agents in a single program

  Incorrect

  The CEO has ruled out a one-step replacement, and AWS guidance positions agents as teammates rather than replacements, introduced through structured change management.

Where competitors have not moved yet, early adoption can set the pace of the market, but only if the AI reaches production and improves what customers value. Investment levels should follow competitive dynamics, not wait for them to be settled by a rival.

• ATrain its own conversational model on past support tickets

  Incorrect

  Training its own model needs machine learning skills and investment the retailer lacks; AWS's scenario for small businesses relies on managed cloud contact center and AI services instead.

• BUse cloud AI contact center services with low upfront cost

  Correct

  AWS's autonomous call center scenario describes how small and medium businesses can use cloud-based contact centers with AI to offer human-like customer experiences without the significant upfront investment of traditional contact centers.

• CLicense an on-premises contact center platform with AI add-ons

  Incorrect

  Traditional contact centers require significant upfront investment and ongoing operational costs, the barrier the retailer cannot afford.

• DUse AI only to suggest replies to its staff during business hours

  Incorrect

  Assisting staff during office hours leaves customers without help at night; the AWS scenario uses AI to serve customers autonomously and escalates to people when needed.

AI can change competitive dynamics by lowering the cost of capabilities that used to require scale. For smaller firms, managed cloud AI services can close service gaps with much larger rivals without a large upfront investment.

• ALet producers' and buyers' agents coordinate on its platform

  Correct

  AWS names network effects, value that grows through multi-agent ecosystems where coordination compounds utility, as a long-term growth opportunity; every participant whose agent joins adds value for all the others.

• BBuild a stronger buyer assistant than the rival's on a larger model

  Incorrect

  A better single assistant is the kind of advantage the rival can close with its next release, and its value does not grow as more participants join.

• CCut commission fees using the savings from AI automation

  Incorrect

  Efficiency savings are an immediate gain, but a fee cut is easy for the rival to match and does not compound as the marketplace grows.

• DGive each producer a private assistant inside its own systems

  Incorrect

  Isolated assistants help each producer separately, but without coordination across the marketplace they create no value that compounds with every new participant.

An advantage is most durable when it grows with use. A platform where many parties' agents coordinate becomes more valuable with each new participant, which a rival cannot reproduce by shipping one comparable feature.