Current Edge · Lead story
Silence in notices written for another purpose, six months after the first go-lives, is not evidence of failure.
Picture a programme board asked to approve a first call-off against a national AI framework. The paper opens with £900m. The second slide says the regional scheme covers 70,000 clinicians. The third reports a neighbouring trust's £1.9m contract. This is an illustrative composite, not a reported case, but every figure on those slides is real and in the public record. The difficulty is that none of them says how much AI anyone is using, or what it has achieved. That is the argument of this paper. The public record of health service AI procurement documents purchasing authority with some care. It does not document deployment or outcome. Leaders who size programmes, benchmark suppliers or set benefit baselines from the first set of numbers are using evidence built for a different question. Three kinds of figure dominate the English coverage. They…
SOTA Watch · Morning four
They show that one large vendor is monetising agents quickly.
Salesforce reports that Agentforce, its agent product, reached an annual recurring revenue (ARR) of $1.2 billion in the first quarter of fiscal 2027, up 205% year on year. A quarter later it reported ARR above $1.5 billion, up over 240%, with 7.0 billion agentic work units (AWUs) delivered to date across Agentforce and Slack, 3.2 billion of them in the second quarter. Those are builder-reported figures: company-defined, unaudited key performance indicators, published by the party selling the product. For a portfolio owner deciding how much to commit to agent platforms, the question is what these numbers license. They do not show that its customers are earning a return. The two ARR figures are not like for like. The second-quarter release states: "Effective Q2 FY27, Agentforce ARR includes our AI offerings, Slackbot and Headless 360". Futurum Research, an independent analyst, notes…
Regulation & Assurance · Morning four
None of the orders read for this brief finds that any Workday tool produced a discriminatory result.
Two beliefs circulate among employers that buy screening software. The first is that a regulator has already ruled on algorithmic bias in hiring. The second is that the vendor, not the employer, carries the risk. The public record supports neither. It supports something narrower: one settled federal case about a rule that named ages, one private case against a vendor that has survived the pleadings without reaching the merits, and a federal policy that changes what the regulator is likely to pursue. This is not legal advice. EEOC v. iTutorGroup, No. 1:22-cv-02565, was brought under the Age Discrimination in Employment Act. The EEOC's press release of 11 September 2023 says the company's application software was programmed to "automatically reject female applicants aged 55 or older and male applicants aged 60 or older", and that more than 200 qualified applicants were…
Compute & Energy · Morning four
Do not scale the national 52 GW, or Cleanview's 2 GW of gas, into a figure for data-centre storage.
Anyone planning around storage at data centres needs one number: how much battery capacity sits behind a data-centre meter and is running. None of the sources opened for this brief publishes it. They give a national total, a count of data-centre self-generation that is almost entirely gas, and a short list of individual sites. The figure in between is missing, and the sites that are named differ in status and in the basis of their numbers. The US Energy Information Administration reports that operators had added 8.3 GW of battery storage in the first half of 2026, taking nameplate capacity to nearly 52 GW, up from 43.6 GW at the end of 2025. The source is the EIA-860M generator inventory, which is operator-reported, and the figures are nameplate: what the units are rated for, not what they delivered. EIA also…