Thirteen signals may change DevRel work through 2028. Each section makes a prediction and lists the developments that would support or weaken it.
1. Agent experience as a deliberate practice
By 2027 or earlier, “Agent Experience Engineer” will be a real job title at AI-adjacent infrastructure companies. The discipline of designing APIs, docs, and developer-facing surfaces specifically for AI agents is emerging now; within two years it will be a recognised specialty.
What this looks like:
- API design reviews include “how does an agent use this?” as a section.
- Documentation pipelines have an AI-readability lint step.
- MCP servers and equivalents are standard product surfaces.
- Some DevRel teams reorganise into “human DX” and “agent DX” sub-teams.
2. The unbundling of “DevRel”
The term will continue to fragment. By 2028 it will be plausible that:
- “Developer Advocate” remains the canonical IC title for public-facing advocacy.
- “Developer Educator” stabilises as a separate role with its own career path.
- “Community Manager” stays a distinct discipline.
- “Developer Experience Engineer” becomes a peer engineering role.
- “Developer Marketer” is fully a marketing sub-discipline.
- The umbrella term “DevRel” persists but means less in any specific context.
The prediction follows the existing split between advocacy, education, community, marketing, and developer-experience roles.
3. AI replaces some tactics, not the function
Tactics likely to be substantially augmented or replaced by AI:
- First-line community support. AI assistants handle the “how do I X” questions; humans handle the harder ones.
- Initial draft content production. Posts, social, video scripts.
- Documentation gap analysis.
- Sentiment analysis and community signal detection.
- Personalised outreach at scale.
Tactics that will remain human-led:
- Accountable public voice. Readers can judge signed work against its author’s record; anonymous generated content makes responsibility for its claims harder to establish.
- Strategic relationships. A founder calls another founder; AI can’t.
- Conference speaking. In-room presence matters.
- Community judgement. Hard moderation decisions, conflict resolution.
- Product feedback synthesis at the strategic level.
By 2028, DevRel professionals are likely to use AI for routine production and analysis while retaining responsibility for relationships, judgement, moderation, and product feedback.
4. Open-source and OSS sustainability moves further to the centre
Three developments support this prediction:
- AI training data ethics. Open-source projects whose code is in training data are demanding more visibility on attribution and sustainability.
- Supply-chain security. Companies need to know who maintains the dependencies they ship.
- Founder-burnout patterns. OSS sustainability has visible failure modes; companies that depend on OSS will increasingly fund it.
For DevRel teams, this would make a recurring OSS-support budget with explicit allocation criteria more common.
5. Conferences continue to bifurcate
Two parallel trends:
- Mega-flagships (re:Invent, GTC, Google I/O, Build, GitHub Universe) continue to grow.
- Small, curated, community-led events continue to grow.
In the middle, generic 2,000-person commercial conferences continue to thin.
For DevRel programs, this implies:
- High investment in flagship presence remains worth it for category-leader companies.
- Repeated participation in carefully chosen smaller community events can build continuing relationships.
- Generic conference sponsorship has the worst ROI.
6. The first-touch shifts further toward LLMs
By 2027, more than half of developer initial research will happen through AI assistants rather than search engines, for many product categories. This implies:
- AI-citation optimisation becomes a real discipline.
- Authoritative long-form content rises in relative value.
- SEO content optimised for Google ranking declines in relative value.
- Documentation must be machine-first as well as human-first.
7. Community-CRM tooling consolidates
The community-data category will consolidate into two or three dominant providers by 2027:
- Common Room or a successor as the commercial leader.
- LFX Community Data Platform as the foundation-grade option.
- Custom data warehouses for the largest companies.
Dozens of small tools are unlikely to persist as the long-term shape.
8. The Developer Relations Foundation becomes consequential
If the Foundation succeeds in its 2025 to 2027 phase:
- Professional certifications emerge.
- Standardised job ladders become available for reference.
- Industry compensation surveys carry more weight.
- The Foundation becomes a meaningful voice in industry conversations.
Failure modes are straightforward: governance disputes, lack of resources, founder departure. If those dominate, DevRel remains an informal profession with no central institution. Both outcomes are possible.
9. Hardware-DevRel becomes important again
AI’s hardware dependency is bringing renewed attention to hardware-developer engagement:
- NVIDIA continues to invest in developer engagement around its hardware and software stack.
- AMD, Intel, Google TPU, AWS Trainium / Inferentia all increase investment.
- Edge-AI hardware (Apple Neural Engine, edge inference chips) drives new developer engagement.
Many DevRel professionals who joined the field through cloud / SaaS work will reskill into hardware-and-software-stack work over the next several years.
10. The geographic distribution continues to broaden
The DevRel field has been disproportionately US- and Europe-centric. Possible growth regions include:
- India. Large developer populations, DevRelCon Bengaluru, and a growing cohort of senior Indian DevRel practitioners.
- Latin America. Growing senior cohort, especially in Brazil, Argentina, Mexico, Chile.
- Africa. Particularly Nigeria, Kenya, South Africa, Egypt.
- East Asia beyond Japan. Korea, Vietnam, Indonesia, Singapore continue to mature.
By 2030, the field will look meaningfully more global than today.
11. Compensation continues to align with engineering
Through the rest of the decade, compensation for senior DevRel professionals at developer-product companies will converge with senior engineering compensation. Already, IC4 to IC6 DevRel roles at top developer-product companies command total compensation comparable to staff/principal engineers. This continues.
Implications:
- Career attraction improves.
- Career risk of being in DevRel rather than engineering decreases.
- Engineering-to-DevRel mid-career moves become more common.
12. Some current orthodoxies will be relaxed
Predictions in this category are by nature speculative, but a few patterns worth watching:
- “DevRel doesn’t do sales.” Increasingly, DevRel at PLG companies does contribute directly to revenue. The strict separation softens.
- “DevRel reports to product not marketing.” The strong default has emerged, but hybrid and CEO-direct reporting are increasingly common.
- “Vanity metrics are useless.” True, but they remain useful internally as health indicators if used carefully. The all-or-nothing framing will soften.
13. AI-engineering becomes a permanent sub-field of DevRel
The “AI Engineer” identity, popularised by swyx and others starting in 2023, is recognisable as a category by 2026. It already has dedicated conferences, publications, frameworks, and senior practitioners; its shape in 2030 remains uncertain.
This category includes DevRel at:
- Foundation-model providers.
- AI infrastructure platforms.
- Vector databases.
- Orchestration frameworks.
- Agent frameworks.
- AI-developer tools.
“AI DevRel” is now distinct from “Cloud DevRel” or “API DevRel.”
What could surprise us
Several developments could surprise the field:
- A major AI-product DevRel team failure. A high-profile cut at OpenAI, Anthropic, or a similar company would reshape narratives about the durability of AI-company DevRel.
- A Linux Foundation-level governance event at the DevRel Foundation could either strengthen or undermine the profession’s institutional infrastructure.
- An open-source framework that becomes a standard could change the products and communities DevRel teams support.
- A platform shift (web → AR/VR, or LLM-first interfaces replacing both) could obsolete current channels rapidly.