Market Research Analysts

Impact

Survey platforms such as Qualtrics and SurveyMonkey generate questionnaires, check them for bias and theme thousands of open-text answers automatically. Qualitative work is changing too: Remesh and similar tools run large-scale moderated conversations and cluster the responses, while general models summarise interview transcripts and pull out quotes. The analyst's day shifts from cleaning data and writing first drafts to choosing the question, checking the machine's reading of the evidence and explaining what it means to a client who has to act on it.

Risk

Analysis and reporting are automating fast; human value moves to research design, interpretation and client counsel.

Occupation-level measures place the role in the highest official exposure tier and observed usage of AI assistants is already high, which is consistent with how much of the work is text, tables and synthesis. Questionnaire drafting, open-end coding, cross-tab generation, chart production and first-draft reporting are all being automated in the current window. What remains human is deciding what is worth asking, judging whether the sample and method can bear the weight of the conclusion, catching the plausible but wrong AI summary, and persuading a client to change a decision. Employment in the broader occupation is projected to keep growing, so this is a role being reshaped at speed rather than removed; the analysts most at risk are those whose output is mainly the deck rather than the judgement behind it.

Sector readiness

Very High Adoption Across Research Platforms

The major survey, panel and insight platforms have embedded generative features into the core workflow, so most agencies and in-house teams are already using AI whether or not they have a formal policy. Large research agencies and consumer-goods insight teams are furthest along, including experiments with synthetic respondents; public-sector and academic research moves more cautiously because of methodological and ethical scrutiny.

1. 01

   Design is the job now. Let the platform generate and code; make sure you are the person who frames the research question, chooses the method and defends the sample, because that is what the client is paying for.

2. 02

   Audit the themes. AI-coded open-ends are fast but they merge and miss categories; read a sample of raw responses against every automated theme before it goes into a report.

3. 03

   Learn the limits of synthetic respondents. Clients will ask about them; be able to explain where simulated answers are useful for pre-testing and where they are not a substitute for real people.

4. 04

   Present, do not just deliver. A report that is read aloud by an AI summary has no author; a presentation where you answer hard questions does.

5. 05

   Pick up the quantitative tools. Basic statistics, a bit of Python or R and comfort with data pipelines let you check and extend what the platforms produce.

6. 06

   Stay close to the decision. Ask what the client will do differently as a result of the study, and write the research around that, because studies without a decision attached are the first to be automated away.

1. 01

   Generative features in survey platforms. Qualtrics, SurveyMonkey and their rivals draft questionnaires, flag poor questions and summarise results inside the product, removing whole steps from the workflow.

2. 02

   Automated coding of open text. Language models theme and sentiment-score thousands of verbatim answers in minutes, work that used to take a junior analyst days.

3. 03

   AI-moderated qualitative research. Tools such as Remesh run large live conversations and cluster responses, blurring the line between qualitative depth and quantitative scale.

4. 04

   General-purpose assistants. ChatGPT, Claude and Microsoft 365 Copilot summarise transcripts, draft reports and build charts, and observed usage in this occupation is already high.

5. 05

   Synthetic respondents. Simulated panels built on language models are being tested for concept screening and pre-testing, reducing demand for some early-stage fieldwork.

6. 06

   Client pressure on cost and speed. Buyers expect faster, cheaper studies now that the tools exist, which pushes agencies to automate everything that can be automated.

Research agencies

Agencies compete on price and speed, so the automation of fielding, coding and reporting is most advanced here and junior analyst intakes are shrinking.

In-house consumer-goods and retail insight teams

Teams embedded in the business are shifting towards continuous data and AI summarisation, with the analyst acting as internal consultant to brand and product teams.

Healthcare and pharmaceutical research

Regulatory and ethical requirements keep more human oversight in design and interpretation, slowing the pace of change.

Public sector and social research

Methodological transparency and sensitivity around populations mean AI is used for drafting and coding but rarely for unsupervised analysis.

1. 01

   Research design and sampling. Knowing which method answers which question, and what a sample can and cannot support, is the skill that AI output most depends on; formal training through a research association is worth the time.

2. 02

   Statistical literacy. Being able to check significance, weighting and segmentation yourself lets you catch errors in automated analysis; a short applied statistics course covers most of what is needed.

3. 03

   Prompting and output verification. Learn how to instruct AI tools precisely and how to test their summaries against the raw data, because unchecked AI output is the main new source of error in research.

4. 04

   Data handling in Python or R. Light scripting lets you clean, merge and re-analyse data outside the platform's constraints and makes you useful to data science colleagues.

5. 05

   Storytelling and presentation. Turning findings into a recommendation that an executive will act on is the part of the job furthest from automation; practise by presenting rather than sending.

6. 06

   Commercial understanding. Understanding the client's business model and decision cycle lets you frame research around decisions, which is where the fee justifies itself.

Open-end coding at scaleExample 1

How

An analyst runs several thousand verbatim survey answers through the platform's AI theming, then checks a sample against each theme and merges or splits categories before reporting.

Gain

Qualitative depth is available on quantitative samples without a week of manual coding.

Discussion guide and transcript synthesisExample 2

How

A general assistant drafts a discussion guide from the research brief, then summarises each focus-group transcript with supporting quotes for the analyst to verify.

Gain

Turnaround on qualitative projects falls from weeks to days.

Report first draftExample 3

How

Copilot produces narrative text and charts from the cleaned data tables, and the analyst rewrites the interpretation and recommendation sections.

Gain

Analysts spend their time on what the findings mean rather than on formatting.

CopywritersMore exposed · exposure 76

AI impact

Generative models produce marketing copy directly, so the output of the role is substituted rather than assisted.

Work moves to

Market research analysts keep a methodological role that copywriters lack; the lesson is to stay on the judgement side of the work.

Data ScientistsDifferent skills, growing · exposure 70

AI impact

AI writes much of the code, but the demand for people who can build and validate models on business data keeps rising.

Work moves to

Quantitative modelling and engineering skills; a realistic path for analysts who extend their statistics and scripting.

Brand ManagersComplementary, less exposed · exposure 56

AI impact

Brand managers use AI for analysis and content, but ownership of strategy, budget and stakeholder decisions stays with the human.

Work moves to

Commercial decision-making built on research insight; the natural client-side move for an experienced analyst.

Our report is one reading of the evidence. This section is the other dataset: what people who do or know this work make of it. Nobody has scored this role yet. Sign in to add yours.