A major global retailer isn’t slowing down for market turbulence—it’s ramping up. In this ETR Insights interview, the company’s Head of Technology reveals how a “push ahead, but with guardrails” strategy is fueling aggressive generative AI adoption while protecting customer data. Early wins include cutting support times in half with Salesforce Agentforce, slashing false positives with Darktrace AI, and securing flexible vendor contracts to future-proof investments. The roadmap ahead? Smarter surveys, weather-driven freight routing, and sharper AI pricing strategies.
Vendors Mentioned: CrowdStrike (Charlotte AI, Falcon) / Darktrace / Domo / Elastic / Microsoft (365 Copilot) / OpenAI / Qualtrics / Rubrik / Salesforce (Agentforce/Einstein, Tableau/Tableau Agent) / SentinelOne / ServiceNow / Zendesk
Our guest, an IT executive surveyed alongside roughly 2,000 peers, says their company is pressing ahead with technology spending, with as much as 10% or 15% budget growth. Although the company temporarily halted some big initiatives amid tariff uncertainty, management plans to resume them once trade tensions clarify, leaving the overall investment trajectory firmly upward. “I think it doesn’t matter if there’s a tariff situation. When it comes to generative AI, that’s something that we definitely want to invest heavily in. Not to say that we want to be at the bleeding edge, but at the same time, we try to be careful and evaluate our business objective and also compliance.”
They are pressing ahead, but with guardrails tied to data privacy and ISO compliance. “The sector that we are in, we’re dealing with a lot of consumer data and PII, so we have to be careful about the AI or any tool that we adopt to AI.” Management now budgets for AI in the high six or low seven figures, well above outlays reported by most retail peers. “Sometimes also it has to do with the business revenue, because we do evaluate the quarterly spending and revenue, and then adjust from there.”
Our guest notes. “We’ll probably accelerate at the beginning or spend a large sum of investment, and then once the solution or any AI-related stuff actually settles down, we anticipate it will be steady, but not dramatically increasing 10% a year.”
ETR Data: According to preliminary data from ETR’s JUL25 Macro Views Survey, respondents nearly equally-prioritize investments in AI-powered automation and advanced cybersecurity to enhance operational resilience, in line with the priorities discussed by this global head of technology.
Vendor behavior is already shifting to match; Salesforce and other suppliers are floating contracts that let customers reallocate unused funds to future subscription models. “If there’s anything that we cannot spend [with Salesforce], we can actually change the model, meaning they allocate the fund for the next project or a different model.” Looking out 12 to 24 months, “At the beginning we commit to a fixed amount, and then later on it’s more like a subscription basis.”
At this firm, they use Salesforce [Agentforce] to triage routine customer-support requests, and an AI-powered asset manager to speed image searches in marketing campaigns. “We did a measurement, and based on the time that they spend and the SLA, we see the SLA actually improve – cut in half – and also the time that the CSR is spending on Zendesk or helping customer actually reduces.” Einstein screens routine customer-support queries so agents can eliminate some of the basic questions and focus on tougher problems. With images, “Before, when they created any digital asset, they had to put a lot of tags into the digital image or video. But now, with the AI component, we can search, say, ‘A man holds the apple wearing blue jeans,’ and it still comes up.” Business intelligence tools like Tableau and Domo now give more predictive analytics from the get-go, “So the human can be validated accordingly.”
ETR Data: In ETR’s MAY25 AI Product Series, 58% of respondents say they spent less than $1M on AI in the past 12 months, with 22% spending less than $100K. However, most (87%) anticipate increasing AI spend in the next 12 months. More than a quarter (28%) anticipate increases in the 1-10% range, more than a third (36%) in the 11-20% range, and nearly a quarter anticipate 21% or greater increases. View the full AI Product Series, and vendor reports, on the ETR Platform.
Darktrace’s AI platform has sharply reduced false positives and accelerated incident response; the tool’s dashboards now double as evidence vaults for compliance and insurance. “Without the Darktrace InfoSec, network guys are having a hard time sometimes trying to do tracing of the log, and then trying to see the event or the attack, and coming from which direction. Now, by using the AI/ML model, the guys actually were able to look at the summary level and then pretty much eliminate some of the false threats, create a forensic case, and incident response.”
With regards to other vendor-embedded AI features, this organization uses CrowdStrike Falcon for endpoints, is trialing Rubrik for backups, and likes Elastic’s AI for cleaner analytics. “I think for [Elastic], we see a lot more data accuracy and consistency in the data.” Tableau Agent, though, remains a work in progress. “It has mixed feedback between the finance team and the marketing team, [primarily due to internal differences in data utilization between finance and marketing team], and so not so much an Agent problem. Right now, we’re still trying to sort it out between the cross-functional team and IT.”
To date, CrowdStrike’s outage is not prompting a defection at this enterprise. “The financial impact for sure was there, and the operational impact for sure was there. However, we see it as a one-time incident because it’s not something that happens every couple years.” Our guest intends to stay with CrowdStrike, keeping SentinelOne as a fallback. “We do have a business unit, a subsidiary, using SentinelOne. Worst-case scenario we can swap to SentinelOne, but that’s our Plan B.” Of note, however; our guest adds that any extra charge for CrowdStrike’s Charlotte AI will likely need to be less than 10%, and aligned with user volume. “We’d have to justify it based on the quantity of the users that we have. It’s to be determined, I would say.”
ETR Data: ML/AI spending growth has moderated on a year-over-year basis, but the sector still registers the highest aggregate Net Score by a margin, as seen in preliminary JUL25 Technology Spending Intentions Survey (TSIS) data. While customers are still increasing spend at robust levels, new Adoptions within the sector continue to soften incrementally.
More than half of companies investing in AI admit to ETR that they aren’t tracking payback or haven’t seen any yet. But our guest’s organization is primarily using Salesforce’s Einstein atop Zendesk, where ROI is more clear: 50 agents now handle 5,000 to 7,000 monthly tickets in roughly half the time. “We did a measurement, and based on the time that they spend and the SLA, we see the SLA actually improve, and also the time that the CSR is spending on Zendesk or helping customer actually reduces.”
Our guest is already mapping out roles for agentic AI. First, they want to inject an AI agent into live customer-survey calls, a bot that can add content in real time. “They make it richer content, so the human absorbs the information better and softer, and also is willing to respond back to the question in the Q&A.” They also hope to inject weather-driven routing intelligence into the company’s transportation management system to keep freight moving efficiently. “If a thunderstorm is coming, maybe you can [use AI] take a different route, or things like that.” Products like Salesforce’s Agentforce and Microsoft Copilot showing high attach rates among users; while total spending may dip, our guest is open to sustained investment if the tools deliver clear value, and vendors maintain transparent relationships. “I believe they will change how they are going to be charging us. Once we stabilize, the spending will be less, and then we pull back on the funding, and they still have to maintain the revenue for any AI product or services driving the revenue. But if the vendor is willing to build that kind of relationship, I think it could be a win-win situation.”