Discussion
  • Agency

Which SEO skills are you investing in now, and which are you deliberately letting go because of AI?

Published 1 reply

Discussion starter: an agency question, posted with an answer from Niraj Raut to open the thread. If you have dealt with this on a site, reply with what you saw, especially where it differs.

Disclosure: Niraj Raut, who posted this answer, runs the SEO consultancy linked at the end of it.

1 reply

  • admin
    Admin
    Original poster

    The skills worth investing in are the ones where AI makes judgement more valuable rather than less: diagnosing complex technical problems, designing and reading experiments, understanding how AI search systems retrieve and cite, content strategy built on real information gain, reputation and digital PR, working with engineers and product teams, and governing AI-driven content pipelines so they stay accurate. The skills to let go of are the ones AI now does adequately and quickly: assembling keyword lists by hand, writing commodity copy and meta descriptions at scale, building reports manually and running basic audits by clicking through tools. The shift is from producing outputs to deciding what matters and checking that it is right.

    Skills rising in value

    Skill Why it matters more now
    Technical diagnosis: logs, rendering, crawling, migrations AI can list issues but struggles to decide which one caused a specific drop
    Data analysis and experimentation Separating signal from noise in volatile search data is the core of good decisions
    AI search understanding and measurement New surfaces, new reports and new ways of being cited need people who understand them
    Content strategy with information gain Generic content is abundant; deciding what is worth creating is scarce
    Digital PR and reputation What others say about a brand increasingly shapes both rankings and AI answers
    Engineering literacy (SQL, Python, APIs) Working with data at scale and automating checks
    Stakeholder communication Explaining uncertainty and trade-offs to leadership is harder, and more important, than ever
    Governance of AI pipelines Keeping generated content accurate and on-brand at scale

    Skills declining in value

    • Manual keyword list building and clustering.
    • Writing large volumes of routine copy, such as meta descriptions, product blurbs and basic articles.
    • Assembling reports by copying data between tools.
    • Running audits that consist of exporting tool warnings without prioritisation.
    • Memorising tool interfaces and checklists rather than understanding systems.

    These tasks still need doing, but they no longer justify a person’s time at the rates they once did. The value has moved to the review and decision that follow them.

    How I would structure development

    1. Teach systems, not tools. People who understand how crawling, indexing, ranking and retrieval work adapt when tools change.
    2. Pair junior staff with real diagnosis work instead of production tasks that AI now handles.
    3. Build data skills across the team, at least enough SQL and scripting to query Search Console exports, logs and analytics.
    4. Practise explaining uncertainty, through writing briefs and presenting findings with confidence levels.
    5. Create a verification habit: every AI-assisted output checked against sources before it reaches a client or goes live.

    Hypothetical example:

    An agency used to train new hires by having them write meta descriptions, build keyword lists and compile monthly reports. After adopting AI tools for those tasks, it changed its training: new hires shadow migration projects, learn to query logs and Search Console data, review AI-generated drafts against sources, and present one finding a month to a client with the evidence behind it. Within a year, junior staff contribute to diagnosis and strategy much earlier, and the agency’s work is less exposed to price competition from tool-driven competitors.

    Watch for this:

    Letting go of production skills entirely can leave a team unable to judge AI output. People who have never written a good title or done keyword research by hand struggle to recognise when a tool gets it wrong. Keep enough hands-on practice for the judgement to be real.

    Hiring for the new balance

    Hiring criteria shift accordingly. Portfolios of written content or long lists of tools matter less than evidence of problems diagnosed, experiments designed, and decisions explained. Useful interview exercises include giving a candidate a real traffic drop with Search Console and log extracts and asking for a diagnosis, or giving them an AI-generated audit and asking which findings matter and which are noise. Candidates who can say what they would check next, and why, tend to be the ones who thrive when tools handle the routine work. Curiosity about how systems work, and comfort saying “the evidence is not strong enough yet”, are worth more than familiarity with any specific platform.

    The career implication

    For individuals, the safest position is to be the person who can explain why something happened and what to do about it, with evidence. That combines technical understanding, data skills and communication, and none of those is easily automated. For agencies, it means pricing and staffing around judgement and accountability rather than hours of production.

    Invest in diagnosis, experimentation, AI search understanding, content strategy, reputation work, data skills and communication. Let tools take the production tasks, keep enough hands-on practice to judge their output, and build the team around decisions rather than deliverables.

    Need help with this on your own site?

    Niraj Raut works with businesses in Nepal, Australia, the UK, Europe and the US on technical SEO, ecommerce SEO, local SEO and AI search.

    See SEO services

Add a reply

Replies are open to approved members. Log in or apply to join.