Expert Systems with Applications Submission Process

First decisions are fast (median about 5 days) and weighted toward desk screening. For papers actually sent to reviewers, the realistic range to a post-review decision is often 6 to 9 weeks, while edge cases diverge sharply: scope-mismatched or application-thin papers are returned within days, and domain-misrouted papers can wait longer while the right handling editor is found. Overall, submission to acceptance runs about 147 days on the journal median.

If you want an outside read before you open Editorial Manager, use the free manuscript readiness check to test whether the applied-AI contribution survives a fast desk screen.

Initial Quality Check: completeness and policy fit

The first layer is administrative but still decisive. Before a handling editor reads for scope, the initial quality check verifies authorship and CRediT contributor roles, competing-interest declarations, ethics and consent statements where human or clinical data are involved, a plagiarism and similarity scan, and the data-availability statement, alongside mandatory Elsevier Highlights and usable figures. A submission can look finished in the portal and still be weak if the abstract, Highlights, and first figure do not tell the same applied-AI story.

Editorial assignment: routing by application domain

ESWA routes to a handling editor matched to the application area (healthcare, finance, manufacturing, transportation, cybersecurity, energy, and similar domains) rather than a single central office. The domain you signal in the title, abstract, and keywords determines which editor reads the contribution first, and a misrouted paper can read as narrower or less novel than it is.

Peer review: applied-AI assessment after the desk screen

Manuscripts that clear the desk screen move to two or more reviewers. The reviewer job is not only to check that the method is correct. It is to decide whether the application is novel, whether the baselines are current, and whether validation reflects a realistic decision setting rather than a benchmark in isolation.

Final decision: scope fit stays live after reports return

Even after review, the decision still turns on applied-AI fit. A technically sound paper can be returned if the reports show the contribution is incremental for the application area, the baselines are stale, or the practical payoff is asserted rather than measured.

This is where the short first-decision median comes from. Before any reviewer is assigned, an ESWA handling editor reads the abstract, the mandatory Highlights, and the first figure, and decides whether the paper is a credible applied-AI contribution in scope for the journal.

At this stage the editor is effectively asking:

• does this manuscript solve a real-domain problem, or is it a method looking for an application?

• is the application in scope for ESWA rather than a methods venue such as TPAMI or JMLR?

• is the package complete, with Highlights, a data-availability statement, and figures that match the claims?

Because this screen is fast, a decision that arrives within days is almost always a desk return rather than an acceptance. The turnaround is deliberately quick so authors can re-route to a sister venue without a long wait.

Papers that pass triage go to two or more reviewers, who typically assess:

• application novelty and real-domain contribution

• baseline currency and the strength of the comparison set

• validation in a realistic decision setting, not only benchmark accuracy

• whether the practical consequence of errors is quantified

• clarity of the contribution in the abstract, Highlights, and figures

ESWA uses single-blind peer review, so reviewers see author identities while staying anonymous themselves, and Elsevier's article transfer service lets a handling editor offer a desk-returned manuscript to a sister journal without a fresh upload. Post-review decisions arrive around 62 days after submission on the journal-level median, though a single manuscript can move faster or slower depending on reviewer availability and the application domain.

• Reject (fast, pre-review): a desk return from the handling editor, usually on scope fit, application-thin framing, or an incomplete package. Re-route to a sister venue or revise the framing before resubmitting elsewhere.

• Major revision: substantive reviewer concerns, often about baselines, validation realism, or the strength of the applied claim. The revised paper usually returns to the same reviewers; respond point by point.

• Minor revision: the paper is essentially accepted pending specific fixes. Respond carefully and promptly.

• Accept: uncommon on the first round; usually follows a clean revision.

Four recurring patterns return otherwise-capable ESWA packages in the first decision window:

• Treating a fast first decision as good news. At ESWA a quick decision is almost always a desk return. If the status moves from With Editor to a decision without passing through Under Review, the manuscript was screened out before review, not accelerated through it.
• Signaling the wrong application domain. The title, abstract, and keywords route the paper to a handling editor. A generic framing routes it to the wrong reader and makes the applied contribution look thinner than it is.
• Submitting without mandatory Highlights or a usable data statement. Elsevier integrity checks and the desk screen both expect these. A missing structured element is an avoidable pre-review return.
• Leading with the method and reaching the application last. The desk screen reads the abstract, Highlights, and first figure. If the real-domain problem is not visible there, the paper reads as a method in search of a use case.

Check whether your ESWA abstract makes the applied contribution visible to the desk screen →

Check if your ESWA submission package is complete before the integrity check →

Check whether your manuscript signals the right ESWA application domain for routing →

This guide tells you what ESWA editors look for in the first decision window; the review tells you whether your paper passes that screen. 60-day money-back guarantee; authors retain all rights and we do not train models on submitted manuscripts.

In our pre-submission review work on Expert Systems with Applications submissions, three patterns account for most of the manuscripts that stall in the fast first-decision window, before a reviewer is ever assigned.

The application is buried below the method

We repeatedly see ESWA manuscripts where the abstract and introduction open with the model architecture and only reach the real-domain problem in the final paragraph. Because the handling-editor desk screen reads the abstract, the mandatory Highlights, and the first figure, an application that surfaces late reads as a method in search of a use case. The fix we push is to make the decision the system supports, and the domain it operates in, legible in the first three sentences and in at least one Highlight.

The submission package is incomplete for the integrity check

A related pattern is a manuscript that is scientifically ready but operationally unfinished: Highlights missing or written as a second abstract, a data-availability statement that says available on request rather than naming a repository, or figures that do not match the numbers in the results. ESWA's Editorial Manager integrity check and the handling editor both screen for this before review, and we treat a complete package as a desk-screen prerequisite, not a formatting afterthought.

Baselines and validation read as incremental to a domain editor

The third pattern is a comparison set drawn from older work or the authors' own prior papers, paired with validation that stops at benchmark accuracy rather than a realistic decision setting. A handling editor who works in that application area registers a stale baseline immediately, and it reframes a sound contribution as incremental. We push authors to refresh the baseline table against current ScienceDirect issues and to state the practical consequence of the model's errors in the deployment domain, because that is the difference an applied reviewer is screening for.

ESWA (JIF 7.5, broad applied AI, high throughput) sits among several adjacent venues, and the desk screen is partly a routing decision:

• choose IEEE TPAMI if the contribution is a top-tier ML method or pattern-analysis advance rather than an application

• choose Knowledge-Based Systems for knowledge-based or expert-systems specialist work

• choose Engineering Applications of Artificial Intelligence when the contribution is engineering-application led

• stay with ESWA when the work is a genuine applied-AI contribution in a real domain with current baselines and measured impact

Last verified June 2026 against ESWA's ScienceDirect insights and Elsevier author guidance. Editorial timing medians shift between updates; treat them as planning ranges and confirm the current figures on ScienceDirect before you submit.