# TMRE Insights Executive Retreat, Virtual meet-up 2024: themes of the day An hour online between retreats: insight pulled towards consulting, self-serve data without chaos, and AI judged on better and happier as well as faster. Source: https://audiencestrategies.com/case-studies/tmre-retreat-2024-virtual-meetup.html Under the Chatham House rule: no names, no companies, no session-by-session account. In 2024 the retreat met only online. On Wednesday 24th July 2024 about 15 senior insight leaders, past and invited retreat participants, joined a one-hour video call chaired by the host. Past participants had asked for a chance to connect between retreats, and this was it. Each person introduced themselves, then the group discussed two questions: how has our world changed in the last year, and where is it going? The call was transcribed only so the host could write notes, and ran under the Chatham House rule. These notes are written to keep to that rule: they aim to carry no names, companies or job titles, and the host reviews them before publication. The in-person retreat planned for that October in Orlando, Florida, never took place: the main conference was postponed because of Hurricane Milton, and the rescheduled event dropped the retreat. A short section at the end records what the Orlando invitees had said they wanted to discuss. ## I. Insight is being pulled towards consulting, and participants questioned where it stops The strongest thread was a change in what stakeholders ask for. Data and tools have spread across organisations, so raw findings are easier to get without the insight team. What businesses still want is someone to make sense of it. One participant said stakeholders were asking for "this need for us to kind of take on more of a business consultant role, and kind of evolve from just insights", and found more energy and more pull on that side than in simply delivering findings. That raised a boundary problem. One participant described stakeholders asking the insight team for work that belongs to other specialist teams, and noted that every hour spent consulting may be an hour not spent on core insight work. Another asked where the function's part in applying insight should start and end: interpretation, positioning, messaging, or further still. A third argued that opening a conversation with a new stakeholder at the level of the business problem changes the relationship: "they move to that pretty quickly. And they're much more comfortable and much more interested in having that strategic conversation." One participant offered a hypothesis: that the move towards consulting may be what drives the merging of insight with analytics and customer experience, because together they can tell both the what and the why. ### So what for your team * Decide where your team's contribution stops, before a stakeholder decides for you. Consulting that crowds out core work is a cost as well as a promotion. * Open new relationships at the level of the business problem rather than the data request. ## II. Self-serve data, and who owns the insight Several participants discussed dashboard adoption and stakeholder training. One described sharply divided views within their team about giving stakeholders direct access to data, from worry about what stakeholders would do with it to frustration at running the same basic cut one more time. One participant cautioned against designing access policies around the anticipated behaviour of a single user. Training stakeholders to serve themselves remained a challenge for several contributors. Another described the continuing effort needed to get stakeholders using the dashboards that already exist, and a third the work involved in helping unfamiliar stakeholders understand and use insight at all. AI sharpened an older question about ownership. One participant observed that when a technology team builds an internal AI platform, it may also expect to produce the insight, which reopens arguments about who does what. ### So what for your team * Treat stakeholder training as part of the product, not an afterthought to the dashboard. * Settle who owns insight produced by internal AI tools before the platform is built. ## III. AI: faster, but is it better, and is it happier? AI ran through the whole call, mostly as a working reality rather than a debate. People described using it to reach insight faster as teams got leaner, to start drafting content and ideas, and to skip the blank page on discussion guides and questionnaire outlines. Vendors were rushing AI summaries and search into their platforms, which brought confidentiality reviews with it. One participant heard vendors describe using AI to check that survey respondents are who they say they are. Another had met a company selling entirely synthetic survey responses, claimed to be highly accurate. The claim drew scepticism, although buyers might still pay for such responses. Cautions sat alongside. One participant read signs of a backlash brewing against the hype, and argued the real question is where a tool sits on the path from promise to proof, and when to commit; intellectual-property protection and tool approval were slowing adoption too. Another described a demonstration in which AI made a research project more efficient overall, while the final step, putting the recommendations in front of senior leaders, still worked better done by a person. A third pushed beyond speed and cost: "I'm hoping that my team isn't just looking at faster and cheaper, but we're getting to something better too." The host added a fourth word. Even where AI does not make work quicker, if it makes the work easier and more enjoyable, that is worth having for its own sake and may help keep good people: better, quicker and happier. One participant urged teams to consider how new technologies change consumer behaviour. ### So what for your team * Ask what AI makes better as well as what it makes faster or cheaper. Count easier and more enjoyable work as a real return. * Test where human input improves the final recommendations. * Consider how new technologies change consumer behaviour. ## IV. A changing data landscape, and a wider idea of who research is for One participant argued for reviewing contractual commitments when the reliability of underlying data is changing, rather than locking into tools that may soon be obsolete. Another described companies pushing hard for measurement frameworks that show a return. Inflation shaped the year for one participant, who warned that failing to understand and act on it would turn this year's work into next year's budget-cut conversation. Another said the hard part was balancing what wins now with foresight, across countries with their own economic and political currents. One participant named the political landscape as an elephant in the room for consumer sentiment. The discussion also raised barriers to research participation, including for people with disabilities. ### So what for your team * Test whether every person you want to hear from can actually complete your research. * Review contract flexibility where the underlying data is changing. ## V. Talent: what is a data scientist anyway? The host asked whether data science teams were shrinking or being folded under insight. The answers turned on a definition. One participant said the title has lost its meaning, because analysts now call themselves data scientists while the job is unchanged, which makes it hard to tell whether teams are growing or contracting. Another said the title is now applied to people whose work spans quite different skills. A third was reconsidering which skills a new analytics team needed, data engineering, data science or strategic consulting, before hiring into it. On the market itself, one participant hiring at the time found a wider and stronger field of insight and analytics talent than expected, helped by layoffs in neighbouring sectors. One participant argued that professional judgement would remain important as analytical tools become more widely available. ### So what for your team * Define the jobs before the titles. Decide which mix of engineering, science and consulting you need, then hire for it. * Invest in the judgement needed to interpret and challenge AI-generated findings. ## What the Orlando room planned to discuss This section is pre-event survey material for a retreat that was then cancelled. Before the planned October 2024 retreat in Orlando, invitees answered a survey about the topics that mattered most to them. A summary made at the time, with AI help, grouped their needs into ten areas: * implementing AI and new technology, and balancing it with human expertise; * moving insight from service provider to strategic partner, and showing its return; * turning insight into strategic recommendations; * integrating insight, analytics and data science in one team; * storytelling and communication for different audiences; * attracting, developing and keeping talent, including in hybrid teams; * research methods and data quality; * balancing speed, quality and cost, including when to do research in-house; * foresight and spotting weak signals; * staying human-centred, and making research diverse and inclusive. The list echoes the July call almost point for point. The conversation it was meant for never happened. ## The thread through the hour The host's closing observation was that he had never seen so much nodding on a call. People in very different industries recognised each other's problems: the pull towards consulting, the self-serve dilemma, the meaning of a job title, and the gap between AI's promise and its practice. The invitation was to contact whoever had said the thing that struck them most, and keep talking before meeting in person.