Which audience research tools should a planner automate first? Start with the work clients never see and rarely value: competitor audits, social listening round-ups and first-draft personas. Automate the gathering, keep the thinking. Planners who get that split right end up with more time for the part of strategy nobody can buy off the shelf.
The timing is not academic. WARC's Future of Strategy 2026, its annual survey of mostly agency-side strategists, names AI adoption as the biggest change to the strategy role. 56% of strategists now say AI replacing strategy roles is their top expected threat, up 15 points in a year. Yet 66% still see clear reasons for optimism. Both numbers are right, and the gap between them is what this piece is about.
Why are strategists so worried about AI?
Because a lot of what gets billed as strategy is really assembly. Pulling screenshots of competitor ads. Reading 400 comments to find the three that matter. Building a persona slide from a survey that was out of date before the brief landed. If that is what you think the job is, the fear is rational. Machines are very good at assembly.
WARC's own framing points somewhere more useful. The report places the strategist's value in judgement, in knowing which questions are worth asking and in sensing what people might want next. None of that is assembly. The threat is real for the planner who only compiles. For the one who interprets, it looks more like a promotion.
Why does desk research swallow so much of a planner's week?
Economists use the term market failure for a situation where a market, left to itself, puts resources in the wrong place. Nobody is being stupid. The incentives simply point the wrong way.
Agency strategy has a small, chronic version of this. Clients buy strategy as a fee, usually folded into a retainer or a project. Nobody pays a line item for desk research, so it is never priced, never measured and never questioned. It just expands to fill the evenings. Meanwhile the scarcest resource in the building, a senior planner's judgement, gets spent on copying and pasting.
The result is a system that under-supplies the thing it values most. Clients get decks that are thick with data and thin on a point of view. Planners burn out doing work that does not use their best skills. Everyone loses a little, every week, and nobody sends an invoice for it.
That is why efficiency is not the dull part of the AI story. It is the correction. When the gathering takes minutes instead of days, the hours go back to the work clients actually pay for: the tension worth building on, the proposition, the leap. Efficiency is the mother of innovation, not because faster is better, but because freed time is the only raw material new thinking has.
What should planners automate first?
Rank every recurring task on two tests. How often does it come round? And how much judgement does it genuinely need? High-frequency, low-judgement work goes first. Here is a starting order for a typical agency planning team.
| Planning task | Automate? | What the tool should do | What the planner still owns |
|---|---|---|---|
| Competitor ad and content audit | Yes, first | Collect and categorise a category's ads and organic posts in one pass | Deciding which competitor move actually matters |
| Social listening round-up | Yes | Cluster conversation themes and track how they shift week to week | Spotting the tension worth writing a brief around |
| First-draft personas | Yes, with checks | Group real audience content into draft personas and pain points | Testing them against interviews and client knowledge |
| Trend scanning | Partly | Flag topics that are rising fast | Telling a fad from a genuine shift |
| Creative pre-screening | Partly | Score likely clickability before media spend | Brand fit, distinctiveness and risk |
| The brief, the proposition, the leap | No | Supply evidence on request | All of it |
Notice the bottom row. The brief and the proposition stay human, not out of sentiment but because that is where the commercial value sits. A client can buy a competitor audit anywhere. They hire you for what you do with it.
The top three rows are the job Defy Monitor was built for: live competitor, audience and conversation intelligence on every client and category you work in, so the audit already exists when the brief arrives.
How do you choose audience research tools without buying another dashboard?
Most planning teams already keep a graveyard of logins. Here is a practical way to pick a tool that gets used rather than admired.
- Log two weeks of planner time. Split it into gathering, analysing and thinking. You need your own baseline, not a vendor's.
- Pick one recurring deliverable. The monthly competitor report or the pre-pitch category scan are good candidates. Automate that end to end before touching anything else.
- Prefer real behaviour to stated intent. Tools that read what audiences and competitors actually publish show you what people do. Use surveys to test hypotheses, not to generate them.
- Demand traceability. Every AI conclusion should click through to the posts or ads behind it. If you cannot show a client the evidence, it does not go in the deck.
- Test it on a category you know cold. Run the tool on a client you have worked with for years. When it tells you something you know is wrong, you have learnt how far to trust it.
- Put it where the team already works. A tool your planners can question in plain language, inside an assistant they already use, beats a dashboard that needs a training day.
- Decide in advance where the saved hours go. If you do not, they will vanish into more meetings.
What stays human in strategy?
Judgement, for a start. AI can cluster a million posts into themes. It cannot tell you which theme a cautious CMO will defend in front of her board. It can draft personas from real audience content. It cannot sit in a focus group and notice the silence after a question.
Then there is taste. Strategy is mostly choosing what to leave out. Machines are generous by default and give you everything. The planner's job is to throw most of it away and defend what remains.
And there is the client conversation. The best strategic insight is often the one the client half-knew and needed someone to say out loud. That takes trust, context and nerve. None of those come with a subscription.
What does this mean for planning teams and juniors?
The awkward question is how juniors learn if desk research disappears. Desk research was the apprenticeship. It taught people what a category looks like from the inside.
The answer is to redesign the apprenticeship, not to protect the drudgery. Give juniors the automated output and make their job interrogating it: what is missing, what looks wrong, what is surprising. That builds judgement faster than formatting screenshots ever did. It is the split we designed the Defy platform around: machines gather and sort, people question and decide.
WARC's optimists and pessimists are looking at the same machine. The difference is what they plan to do with the time it hands back.
