AGTNEWS
Can Agentic AI Break the SAP Consulting Cost Trap?
A UK consultancy claims a 59 percent reduction in S/4HANA migration effort. The reason it matters is the 17,000 customers who still haven't moved.
A UK consultancy claims a 59 percent reduction in S/4HANA migration effort. The reason it matters is the 17,000 customers who still haven't moved.
Nearly two-thirds of SAP's installed base still hasn't migrated to S/4HANA, and the reason isn't indecision. It's the bill. A UK consultancy thinks agentic AI can change that math, and is putting a number on the bet: a 59 percent cost-weighted reduction in programme effort, claimed in a paid press release placement carried by USA Today on May 23. Whether that figure is real, replicable, or marketing is the question that matters for the roughly 17,000 customers Gartner projects will still be on ECC when mainstream support ends in 2027.
The structural pressure is not in dispute. According to CIO.com, citing Gartner analyst Fabio Di Capua, only about 39 percent of the 35,000 SAP ECC customers eligible to move had completed a migration to S/4HANA by the end of 2024. Gartner projects that nearly half of those customers will still be on ECC when mainstream support ends in 2027, with around 13,000 customers still on the legacy platform in 2030. IDC, via analyst Mickey North Rizza, pegs the holdout share at 40 to 45 percent through 2027. The numbers are projections, not measured outcomes, and the analysts differ on the slope, but the picture is consistent: a majority of the install base has decided that the cost of moving exceeds the cost of staying.
That cost is the point. Gartner client work cited in the same CIO.com piece puts individual S/4HANA programmes in a $2 million to $1 billion range, with three-to-seven-year timelines as typical. SAP's own historical delivery record, cited in the Deshoring Manifesto that Resulting IT published to frame its launch, has been that roughly half of projects deliver on the original business objectives and around 70 percent run over budget or schedule. The 2027 cliff is real, but SAP has also shipped a "private edition, transition option" that extends some large customers to 2033, narrowing the binary-deadline frame vendors like to use.
The vendor behind the 59 percent claim is Resulting IT, a Warrington-based independent SAP consultancy run by CEO Stuart Browne. On May 23 the firm announced two products, S4SensAI and ABAPBanZAI, built around specialist AI agents that take on roles in a typical S/4HANA delivery team: architect, programme manager, business case analyst, development lead. The agents produce plans, RAID logs, RACI models, business cases, and architecture guidance for RISE, GROW, and ECC migration projects. The 59 percent figure, in the company's framing, is a blended programme-level number for cost-weighted effort reduction. The S4SensAI product page carries a separate, larger claim of "up to 90 percent faster and 95 percent cheaper" against traditional SI delivery. Those are different measurements: 59 percent is the blended programme figure, 90/95 is the per-task claim. Conflating them flatters the headline.
Browne's framing of the play is "Deshoring," a term he coined for the consultancy's own marketing and laid out in the same-named manifesto and in a LinkedIn article he published. The argument is that the offshore cost arbitrage that defined the last two decades of SAP delivery has compressed, from roughly a 5:1 cost ratio in 2006 to about 2:1 in 2025, while the productivity gap has not closed. Deshoring, in his telling, replaces offshore consulting hours with agentic AI workers, paired with a customer's existing in-house team. It is a rebrand of the agentic-AI-consulting pitch, but the label lets Resulting IT position against both onshore Big Four and offshore Indian SI incumbents at once.
That positioning matters because SAP itself has an agentic story. The company has been pushing Joule, its AI copilot, and RISE with SAP, its managed-migration programme, as the answer to exactly the cost-and-duration problem that has stalled the install base. Gartner's Di Capua, in the CIO.com piece, framed the migration backlog as the most pressing issue customers face. If a single UK independent can credibly claim a 59 percent cost-weighted reduction, the question is what that does to the competitive landscape between vendors, Big Four consultancies, and SAP's own RISE channel.
The honest answer is that nobody outside Resulting IT has yet published numbers. No independent benchmark of S4SensAI or ABAPBanZAI was available at launch, no named customer program was disclosed behind the 59 percent figure, and the ROI methodology is not public. The 59 percent is one vendor's own analysis, published in a paid press release placement, on a product that has been on the market for two weeks. The same caveats that apply to any single-vendor benchmark apply here, with extra weight because the consultancy and the product come from the same source.
There is also a real question about whether the cost ceiling, not the cost ratio, is the binding constraint. If a $2 million to $1 billion programme can be cut to $850,000 to $425 million, that helps. It does not change the fact that a CIO has to defend a nine-figure modernization programme against a working ECC system that still has support, and that the political and operational risk of a failed migration is what has frozen the holdout cohort. Resulting IT's pitch is that agentic delivery reduces the duration and the consulting dependency, both of which would compress risk. The pitch is plausible. The proof is not yet on the table.
What to watch: whether Resulting IT, or any of the larger SIs building their own agentic delivery stacks, publishes a third-party-validated case study with a named customer and a defensible methodology. Gartner's next ERP migration survey, expected later this year, will tighten the 39 percent figure and the 17,000 holdout count. SAP's 2027 mainstream-support cutover, and the take-up of the 2033 private-edition extension, will show whether customers are choosing to ride the cliff or jump before it. The market has been frozen because the economics have not justified the move. Whether agentic AI actually moves that math is the test the next twelve months will run.