What's Changing in the Workforce Holding Your Moat
Your senior engineers are taking calls. Your domain knowledge lives in three heads. Your 2030 architecture bench is being shaped by decisions you were not invited to. A briefing for Chief Human Resources Officers at mid-market European vertical software companies.
Monday, 9:14 a.m.
Every CHRO at a mid-sized European vertical software firm has had the same Monday. Three things in the inbox. Your VP Engineering forwarded Friday's resignation note from your principal engineer, the one who has been with the firm eight years, the one who carries the why-this-data-model decisions across three of your largest customer segments. He is going to a Series A AI-native competitor in your exact vertical. Your CFO attached the engagement-survey report to Thursday's GMT calendar invite with a single-line note: "engineering thriving down twelve points, need your read before this lands with the board." Your CTO forwarded a thread from the architecture channel where two senior engineers are debating whether to ignore the firm's Cursor-restriction policy because they are losing days a week to manual scaffolding.
You have held this seat for four years. The pattern used to be simple. Hire engineers from good schools. Pay them well. Promote the technical leadership pipeline. Protect the codebase and the customer base. The business was understandable.
Something changed. The technical barrier to building a competing product collapsed in 2024. AI-native entrants are reaching feature parity in your vertical on capital that would have been a marketing budget five years ago. Your A-players are reading every AI tooling decision as a signal about whether the firm is being redesigned for them or around them. Your principal engineer who resigned on Friday was not the first warning. Your engagement signal dropped six months ago and nobody reconstructed the engineering-org-level story. Your CTO has more evidence of shadow Cursor use than sanctioned use.
You do not have a recruiting problem. You have an encoding problem. Your firm's workforce strategy is being written every week in decisions you were not invited to. The question is not whether to have one. It is whose handwriting the one you already have is in. Your A-players read it last quarter. Your AI-native-competitor recruiters are decoding it right now. Your 2030 senior bench is being shaped by decisions that look like tooling-policy items and are actually architecture.
This briefing covers what is happening in your sector, what it means for the people function you run, and the three questions your leadership team should answer before the next GMT.
Vertical software just lost the moat that ran it for a decade
For a decade, the economic architecture of vertical software ran on three pillars. Deep domain knowledge encoded in product. Long integration timelines that punished switching. Workflow embeddedness that made the platform indispensable to the customer's daily work. Your firm earned its valuation on those three pillars. AI collapsed the first two in eighteen months, and the third is weakening as AI-native entrants invest in migration tooling that was not economically viable five years ago.
The AI-native entrants in your category are real and funded. Unity Advisory and DiligenceSquared in services-tech. Unlimited Industries, Buildots, and PermitFlow in construction. Procore announced an AI-agent layer that compresses days of customer work to hours. Every vertical has its version. They are not experimenting. They are reaching feature parity in twelve to eighteen months instead of five years, on capital structures that did not exist when your firm was founded. They do not have your customers, your regulatory edge cases, or your accumulated understanding of why the data model works the way it does. They also do not have your technical debt, your legacy customer commitments, or your founding-team-knowledge dependency. They are racing your moat, and the moat is not the codebase.
Your moat was never the code. The code was always replicable. The moat was the accumulated understanding of the domain, and that understanding lives in three heads, and the three heads are taking calls.
Only four percent of Nordic companies achieve strong AI returns. Forty to fifty percent of Nordic AI spend goes to off-the-shelf productivity tools. Your firm is probably in the same shape. The capital is going to Cursor licences and a Copilot rollout. The capital that would actually defend the moat is not on any line in this year's budget. Funded encoding of the institutional knowledge before the senior engineers leave costs less than one founding-engineer departure and produces a successor in two quarters. The CHRO who sees this first writes a different memo. The CHRO who does not watches the moat walk out the door one founding engineer at a time.
Your people function is feeling it every day
Your talent acquisition. Your principal engineer resigned on Friday. Your best lateral hire of the last two quarters came from an AI-native competitor in your vertical who read your firm's hesitation as opportunity. Your graduate offer-accept rate is holding, but the best graduates are asking in second-round interviews what your firm's AI-tooling policy is, and "we are evaluating Cursor" is not an answer that wins them. A-players do not quit on the spot. They start looking quietly. By the time retention moves in the dashboard, they have already decided.
Your L&D. Your firm's L&D budget is allocated to compliance modules, conference attendance, and Cursor-rollout enablement videos. None of it is designed for continuous in-the-codebase coaching, agent-paired pairing on real architectural decisions, or deliberate cohort rotation across customer segments. Two populations are forming inside your engineering org whether you designed it or not. The top decile is using agents daily and shipping features in days that used to take sprints. The rest are practicing the workflow your firm hired them for in 2022. Both groups are becoming who they practice being.
Your compensation architecture. Your engineering rate card and senior-track compensation reward tenure and individual feature output. Agents are compressing the per-feature time. Your senior engineers see the math earlier than your committee does. With AI-native competitors offering thirty percent more, full architectural control, and the chance to ship the rebuild your firm has talked about for two years, the senior engineers carrying your moat are reading every quarter's compensation run for a signal about whether the firm is being redesigned around them or without them. The comp committee that met last month did not see this.
Your employee relations and works council. Your works council, where you have one, has not opened framework-agreement conversations on your AI tooling rollout because nobody told them it was a workforce decision. The regulatory load on people functions has stacked: AI-literacy documentation, high-risk recruiting and performance-management systems, framework agreements on agent-paired work, data-minimisation in customer-service tooling. None of it shows up on the cost-reduction deck on the CFO's desk. All of it lands on you.
Your workforce planning and succession. Your senior bench has six names on it whose departure would degrade the product roadmap by a quarter and the architectural decisions by a year. Three of them came through code-review apprenticeship work that agents have already absorbed. You have not replaced the pathway. The external senior hire in 2030 will cost a thirty to forty percent premium, take twelve to eighteen months to integrate, and bring the previous firm's context. On a long enough horizon, the CHRO who funded a Cursor rollout instead of an encoding pathway in 2026 is the CHRO the board blames for the 2031 architecture opening with no internal candidate.
The apprenticeship clock was always the architecture
Your sector is repricing because the technical barrier collapsed and the moat moved. Your people function is stretched because every tooling decision, every comp committee, every framework agreement is a fragment of a workforce strategy that somebody is writing and you are not always in the room when they do. Same force, two altitudes. But there is a third thing happening that most mid-market vertical software CHROs have not named yet.
Your firm's people strategy is being written every week. The Cursor-restriction policy engineering is ignoring. The architecture-review backlog teaching juniors what the firm does not value. The recruiter call your principal engineer took on Tuesday that nobody flagged. Your A-players read the handwriting last quarter. Your AI-native-competitor recruiters are decoding it right now. Your 2030 senior bench is being shaped by decisions that look like policy items and are actually architecture.
Bell Labs produced the transistor in thirty months and information theory in one paper because the junior sat next to the senior on the hardest problem in the building, not in a training cohort down the hall. Juniors watched every architectural decision in real time. Chemists sat with physicists sat with mathematicians. Graduation was earned on the work, not by time in seat. The apprenticeship clock was the architecture. That architecture is available to your firm for the first time in two centuries, because agents make it affordable to put junior engineers next to senior architects on the codebase areas where institutional knowledge lives, instead of on greenfield side projects where it does not.
You can tell which camp a firm is in from a single question. Of the three architectural decisions made in the last quarter, how many would survive the principal engineer leaving tomorrow? Most mid-market vertical SaaS firms cannot answer. The ones that can have already built the encoding pathway.
You have twelve to eighteen months to author the answer deliberately, or the AI-tooling-policy memo on the CTO's desk writes it for you. Same technology. Same workforce. Different outcome depending on whose handwriting the answer is in.
What most firms are doing. What the winners are doing instead.
Every CHRO has had the AI conversation. Most are making the same four decisions. A smaller group is making different ones.
| The default path | The winners' path |
|---|---|
| Restrict AI tooling for security and IP protectionEngineers use it anyway on personal accounts. A-players read restriction as mistrust and start looking quietly. | Publish three rules that defend real trustRouting by code sensitivity, not by tool. Architecture-review escalation named. A-players watch when an experiment fails. |
| Fund a Cursor rollout and call it capabilityTooling licence count rises. Median engineer output does not. Two populations form inside the engineering org by default. | Ship one architectural role redesigned as a triad unitSenior architect plus mid-level plus agent ensemble as one delivery unit. Junior sits inside every architectural call from week one. |
| Announce firm-wide AI strategy at the all-handsSenior engineers whose roles get redefined have the power to leave. Three of them do. Rebuild stalls. | Six senior engineers author their own role redesigns firstEight-week council cadence. They bring it to the engineering all-hands as work they authored. Defection risk converts to ownership. |
| Run the annual engagement survey as your main HR deliverableYour cadence is slower than the CFO's quarterly tooling-spend review. Cost-reduction writes the workforce answer. You inherit the damage. | Publish the encoding answer in first 100 daysTwo pages, co-signed by CEO and CTO, referenced every quarter. Every AI deployment gets decoded against it. |
Same workforce. Same twelve months. Opposite outcomes. The difference is whose hand is on the pen.
Talent Gravity. Capability Compounding. Succession Readiness.
Three questions every CHRO is tracking. None of them used to be the same question. They are now.
Are your senior engineers staying because of you, or despite you?
Your principal engineer resigned on Friday to an AI-native competitor in your vertical. Your best lateral recruit last quarter came from the same place. Your graduate programme is holding, but the best graduates are asking what your AI-tooling policy is in second-round interviews. A-players watch which way you lean the first time something goes wrong. The ones who start looking quietly have already decided.
Is your median engineer's output materially higher in December than January?
Two populations are forming in every engineering org. The top decile is using agents daily and shipping features in days that used to take sprints. The rest are practicing the workflow they were hired for in 2022. Your L&D budget funds compliance modules and Cursor enablement videos. Neither moves the median. The firms pulling ahead are making the 90th-percentile workflow the firm's default workflow.
Of your six senior names, how many have a successor who carries the why?
Your single-point-of-failure list has six names. At least three of them came through code-review work that agents have already absorbed. You have not replaced the pathway. The external senior hire in 2030 costs a thirty to forty percent premium, takes twelve to eighteen months to ramp, and brings the previous firm's context. The encoding architecture gets rebuilt now or gets inherited as a gap.
What the firms winning this play are actually doing
Four concrete moves any CHRO can start this quarter. Each addresses one sub-function. Pulled together, they become one answer the firm compounds on.
Talent: publish the encoding answer in your first 100 days
Two pages, co-signed by the CEO and CTO. What this firm believes about the institutional-versus-personal question. Which architectural decisions get encoded, by whom, on what cadence. What investments are being made in senior engineers becoming more capable rather than more redundant. What happens to a senior engineer whose role changes when the encoding lands. Referenced every quarter. Every AI tooling decision after that gets decoded against this memo.
L&D: ship one architectural role redesigned as a triad unit this quarter
Senior architect, mid-level engineer, and agent ensemble as one inseparable delivery unit on the codebase area where institutional knowledge is most concentrated. The mid-level does not write tickets while the senior reviews. The mid-level operates the agent ensemble in front of the senior on real architectural decisions from week one. The senior watches the calls in real time and intervenes at the load-bearing moments. Exposure-to-architectural-judgment hours go from eighty a year to twelve hundred. Encoding compresses from years-of-osmosis to two-quarters-of-deliberate-pairing.
Succession: six senior engineers author their own role redesigns first
Do not ship a firm-wide engineering-redesign at the all-hands. That announcement loses three senior engineers in the next quarter because the engineers being asked to support it are the ones whose roles get compressed. Invert. Assemble six senior engineers as the authoring council. The ones whose codebase ownership is broadest and whose departure would hurt most. Their first assignment: redesign their own role at the task level in eight weeks. They bring it to the engineering all-hands as work they authored. The most likely defectors become the most likely defenders.
Trust: publish three rules, one escalation, one red line
Three rules about what engineers can do with agents without asking. One rule about what requires architecture review before merge regardless of whether agent or human authored the code. One rule about what never goes to an agent. Customer PII without sanitisation. Regulatory-edge-case decisions. Security-architecture changes. Published. Defended to the board when an experiment fails. Your A-players watch which way you lean the first time something goes wrong.
The only firm that joins workforce authorship to working infrastructure
The vertical-software CHRO has four places to go. None can do what we do.
The HR-tech vendors (Workday, SuccessFactors, Eightfold, Gloat, Phenom, Personio, Lattice) sell platform software. The data encoding which engineers carry which architectural knowledge IS your moat in machine-readable form. Building on the vendor stack licenses your moat back to the vendor.
The engineering-talent benchmark and search firms (Heidrick Tech, Russell Reynolds Tech, Riviera Partners) run survey-benchmark models. The benchmark tells you the median senior-engineer comp in your geography. It cannot tell you which of your three founding engineers is taking calls right now.
The Big Four and MBB people-strategy practices (Deloitte, Mercer, Korn Ferry, McKinsey, BCG) sell partner-led diagnostics with reusable IP. They will write the strategy deck. They cannot redesign a senior engineer's role at the task level with the incumbent in the room and ship the redesign inside a quarter.
The L&D platforms and AI-tooling certification (LinkedIn Learning, Coursera, Sana Labs, Cursor enablement) sell content libraries. The senior-engineering pathway is a pattern-exposure problem, not a content-library problem. A platform that sells modules cannot rebuild the architecture agents have collapsed.
The market offers four things you have already said no to. GRAIL is the fifth option.
- Senior advisory that speaks GMT, not HR-deck
- Working infrastructure your team uses the same week
- One architectural role redesigned with the incumbent in the room
- An encoding memo CEO and CTO co-sign in your first 100 days
- No vendor lock-in at any phase of the engagement
- Founded by an operator who has carried CPO seats
- Business first, technology second
- Three weeks from first meeting to working infrastructure
- Your team keeps the infrastructure, you own the IP
- Every next step earned by what just shipped, never assumed
Three sessions with your people leadership. Working infrastructure from week one.
We start with your people leadership team because the change cannot outrun the team's personal understanding of what is possible. Three workshops. Each delivers standalone value. Each earns the right to the next.
How to think
Walk through what is happening in your sector and inside your people function. Draft the first page of your encoding memo in the room, with your CEO and CTO present. Your leadership team leaves with a shared view of what the firm believes about institutional-versus-personal knowledge and the first draft of the document the CEO and CTO will co-sign.
Your people infrastructure
A half-day build. Your Head of Talent, Head of L&D, Head of Compensation, and Head of Employee Relations each leave with working infrastructure tuned to their function. Task-level role-mapping agent for senior engineers. Architect-Council facilitation architecture. Framework-agreement drafting for the works council. Three-rules trust document for AI tooling. Infrastructure your team is using the same week.
Ready to run
Refined infrastructure based on real use. The first architectural role redesigned as a triad unit with the incumbent in the room. The published encoding memo, co-signed. The Architect-Council kickoff agenda. The CFO conversation with numbers your CEO can check: external-hire premium math, encoding-pathway forecast, single-point-of-failure-list count.
Three weeks. No IT project. No board vote needed to start. The infrastructure your people function is already running on when we finish.
For your next GMT
The questions where your people leadership cannot agree on an answer are the ones worth an hour on the agenda. Ask these honestly.
If you asked three of your best engineers privately tomorrow whether the firm's AI-tooling policy is being designed for them or to them, what would they say? Do you know the names, and can you describe the evidence for what they would say?
Walk me through one role in your firm that has been redesigned at the task level in the last twelve months. Who did the redesign, what did it change, and what does the incumbent do with their week that they did not do before?
Of the six names on your single-point-of-failure list, how many have a successor inside the firm who could carry the why-this-data-model knowledge by 2027? What is replacing the pathway for the ones whose pathway is gone?
Your workforce strategy is being written this quarter in tooling-policy memos, comp-committee outcomes, and architecture-review backlogs. How many of those fragments are you the author of, and how many are you decoding after they ship?
Of your senior engineers whose codebase ownership is broadest and whose departure would hurt most, have any of them been asked to author the redesign yet, or are they waiting to be presented with a policy on yours?
If any of these produces a disagreement in your people leadership, that is the conversation worth having before the board meeting.
Continue reading
Sources and limitations
This briefing draws on GRAIL's engagement experience with leadership teams and on public company disclosures. Figures are illustrative unless a source is named beside them. A sourced edition is in preparation.