Speed, specialisation and strategic advantage in the AI era

The free-text comments add perspectives rather than explanations for those percentages. Some respondents favor agility or continuing reinvention; one explicitly values both speed and domain expertise; another worries that short-term focus can displace long-term commitment and investment. Read together, the comments reinforce the absence of a simple one-variable answer.

One comment adds a useful reminder that some forms of advantage accumulate on a different timeline:

“What is our speed for?”

The comment brings a different timeline into the discussion. Trust and relationships may depend on repeated interaction rather than a release schedule. That makes “what is our speed for?” a useful leadership question: what should faster learning or delivery help the business understand, improve, or retain? Our survey did not compare these accumulation times or test whether trust protects a business from displacement.

The implication for leaders is a resource-allocation question: which assumptions need to be tested quickly, and which forms of expertise or customer knowledge deserve sustained investment? Faster activity should have a stated purpose — something the business intends to understand, improve, or build — rather than being treated as an end in itself.

Questions the findings raise for leaders

The following questions translate the evidence into an inquiry for the business, not a prescribed strategy. Buyer access, customer value, economics, implementation, and dependencies are familiar disciplines. Their relevance here is to reconsider what each contributes when the core offering may be easier to reproduce and competitors may find different routes to the customer.

If competitors can replicate our offering, what else would they need to win customers — and which of our existing advantages could they bypass rather than reproduce?

The replication estimates make this the first question to ask. Winning requires more than building something similar, and a rival may not need to reproduce every incumbent advantage if it can reach or serve customers differently. Our survey does not establish the conditions for displacement or show that AI has already made an alternative route viable; it makes those possibilities important to test in the specific business.

What has to be true for customers to buy and keep using our offering — in terms of access, demonstrated value, understanding, viable economics, and implementation?

The comments raise the right buyer, evidence of value, market education, the difference between need and business viability, and what customer work can leave behind. Ask what an engagement teaches, what evidence it produces, and what can responsibly be reused. Those are propositions to test in the business, not universal success factors established by this survey.

Which parts of our advantage depend on deep specialization, which must adapt quickly — and what is our speed for?

The divided judgments make “specialization or speed” too coarse a company-wide choice. Identify where expertise, accumulated knowledge, trust, or technical depth creates value and where routines, products, or assumptions need to change. Then ask what faster work should build: better customer understanding, stronger evidence, more effective integration, or capabilities the business can retain. The aim is purposeful adaptation; our survey does not establish an optimal mix or prove that speed produces compounding advantage.

Which dependencies — on technology providers, funding, major customers, cyber resilience, regulation, talent, or infrastructure — could weaken the business even if customers still value our offering?

The two risk questions show why an attractive offering is not enough. A business can retain customer value and still be weakened by concentration, dependency, or external shocks. The relevant task is to identify which dependencies are material in the specific business and how they would be detected, tested, and mitigated.

A snapshot of competition during rapid technological change

The responses capture judgments made from 5 June to 23 August 2026. They are a record of that period, not a permanent account of how competitive advantage works. One respondent noted that newer AI versions had already made earlier assessments of possible applications feel inadequate. That comment does not establish an objective rate of change, but it captures why a time stamp is useful in research on fast-moving technology.

When interpreting the snapshot, it helps to separate three things: changes in AI capability; changes in the cost and practicality of applying those capabilities; and the business consequences that may follow. A capability becoming technically possible is not the same as its becoming economical or reliable in a particular setting, and neither automatically determines how customers, products, business models, competitive structures, or company performance will change.

Those business consequences may also be indirect. A company may encounter them through different choices made by a customer, supplier, or competitor rather than through its own adoption of AI. Some consequences may appear quickly; others may take years to work through products, customer behavior, business models, and company performance.

Our survey does not provide a fixed timetable for those effects or a forecast of when displacement will occur. Its value as a time-stamped snapshot is different: it creates a record that can be revisited to see which practitioner judgments remained useful, which assumptions proved durable, and which were overtaken by events.

Questions this research leaves open

Further research could make several distinctions sharper. First, matched questions could compare the replication horizon of a respondent’s own offering with comparable offerings in the market, using the same scenario and time bands. That would permit a comparison the present questionnaire cannot make.

Second, future work could revisit these judgments prospectively and distinguish estimated replication from observed copying, entry, customer acquisition, switching, and displacement. A broader response set, including more investors and later-stage businesses, could test whether the strongest founder/operator contrast recurs and whether it survives comparable selection-count checks.

Third, the bypass hypothesis could be tested directly: can a competitor serve the same customer need through a different business model, channel, or cost structure without reproducing the incumbent’s assets? Further research could also distinguish capabilities an organization can acquire from those it develops through experience, and examine whether customer work actually produces reusable evidence, data, integration knowledge, or relationships. Acquired and accumulated advantages are useful concepts to investigate, not a taxonomy validated by this survey.

These are open questions suggested by the findings, not a commitment to a future study or a forecast of its results.

Download the full report below.