Prompt engineering is redefining software value
For decades, competitive advantage in software was assumed to live strictly within the lines of source code. The more sophisticated your codebase, the bigger your edge. Today, modern AI tooling is reshaping that advantage, but instead of erasing the value of code, it is expanding the skills required to produce it. IT leaders must recognise that their team’s IP isn’t shifting away from the source code, but rather extending beyond it – requiring engineers to master new capabilities across AI-driven workflows, prompts, and custom instruction sets to truly unlock their codebase’s potential.
In 2025, the global prompt engineering market size was valued at USD 505.43 million and the market is projected to reach USD 6703.84 million by 2034. However, rather than creating millions of “Prompt Engineer” job titles, leadership at software development specialist, Global Kinetic, say the industry is rapidly baking prompt engineering skills into mainstream developer roles, tools, and hiring criteria.
Dewald Mienie, head of architecture and technology at Global Kinetic says company teams are building a deliberate process around AI assistants rather than treating them as ad‑hoc tools.
“The differentiator is less the raw code and more the way teams have encoded their experience into reusable guidance that the AI follows,” he says.
Mienie describes two pillars of what the business refers to as the “Global Kinetic way”. The first pillar is the workflow that defines how teams use AI end‑to‑end, and second is the prompts that embed their accumulated expertise into those workflows. “It’s in this second one, where we have distilled our experience into the prompts, that is where the IP now resides,” he says.
The same principle applies when building a custom GPT and while the model itself may be a commodity, Mienie says the way developers configure it, including what instructions, examples and rules you load into it, will be highly specific to each organisation.
“Those prompts bundle domain knowledge, regulatory nuance, performance expectations and your house style for engineering into a single asset that can drive many different codebases,” he adds.
Rethinking skills in an age of prompt-led IP
Unsurprisingly, this shift is having a significant impact on how IT leaders are hiring and managing teams.
Last year the demand for prompt engineering skills was surging. According to LinkedIn data, there was a 250% increase in job postings for roles related to prompt engineering, with no sign of it slowing in 2026. However, Dippenaar and his senior team are clear that maintaining enterprise standards demands seasoned coders.
“If prompts and workflows are the new IP, you cannot treat prompt engineering as a junior, copy‑and‑paste task. A good prompt engineer must be a seasoned software professional, somebody who has the experience to actually build enterprise-ready code from scratch, without the use of AI,” says Martin Dippenaar, CEO of Global Kinetic. “That depth matters because prompts now encode decisions about architecture, security, resilience and maintainability. Someone who has lived through real production incidents and large‑scale systems knows what to emphasise, what to forbid and where shortcuts will hurt you later,” he says.
To meet the changes, he says the company is training engineers to break work into small, auditable steps, to know where in the software development life cycle AI genuinely helps, and how to encode patterns and constraints into prompts so the AI stays within enterprise guardrails.
Designing your organisation around the new construct
Dippenaar stresses that protecting enterprise quality in this environment means designing an organisation around these new realities.
Firstly, he says companies require senior, T‑shaped engineers who can act as prompt architects and workflow designers, not just individual contributors. Their job is to define the instruction sets, guardrails and review steps that keep AI‑generated code aligned with enterprise standards.
Second, development teams need structured apprenticeship for juniors. Rather than leaving them alone with powerful tools, seniors must teach them how to question and correct AI output, how to understand the workflows and prompts they are working within, and when to escalate decisions.
Revise how you see your IP portfolio
Most importantly, the shift must change how leaders think about their IP portfolio. In a prompt‑led world, Mienie believes the unique value of an engineering organisation sits in a constellation of assets.
“Source code still matters, particularly where it embeds complex algorithms or domain logic, but it is increasingly an output of higher‑order artefacts rather than the primary store of know‑how,” Mienie adds.
Global Kinetic leadership are adamant that AI has not made software engineers or their expertise obsolete, but say It has moved the point of leverage.
“Your competitive edge no longer depends on owning a secret codebase. It’s about turning your best engineers’ judgement into workflows, prompts and custom instruction sets that reliably produce enterprise‑grade software. It’s up to leaders to manage this shift through hiring, training, governance and investment in these new kinds of assets, which will remain a deeply human competitive advantage,” Dippenaar says.
