diff --git a/app/jobs/file_sets_reprocess_job.rb b/app/jobs/file_sets_reprocess_job.rb
new file mode 100644
index 00000000..f0610930
--- /dev/null
+++ b/app/jobs/file_sets_reprocess_job.rb
@@ -0,0 +1,130 @@
+# frozen_string_literal: true
+
+##
+# This job is responsible for finding file sets that may need re-processing and then dispatching new
+# jobs to perform that processing.
+#
+# The reasons are two fold, and addressed by the two jobs:
+#
+# 1. We did not successfully split a PDF; handled by ConditionallyResplitFileSetJob
+# 2. We did not successfully attach a PDF; handled by ConditionallyResplitFileSetJob
+class FileSetsReprocessJob < ApplicationJob
+  ##
+  # @param cname [String, Symbol] when given :all, submit one {FileSetsReprocessJob} per tenant.
+  #        Otherwise, switch to the given tenant and submit a {FileSetsReprocessJob}
+  def self.for_tenant(cname = :all)
+    if cname == :all
+      Account.all.each do |account|
+        account.switch!
+        FileSetsReprocessJob.perform_later
+      end
+    else
+      Account.switch!(cname)
+      FileSetsReprocessJob.perform_later
+    end
+  end
+
+  class_attribute :solr_page_size, default: 1000
+  class_attribute :solr_q_parameter,
+                  default: "(mime_type_ssi:application/pdf OR label_ssi:*.pdf) AND has_model_ssim:FileSet"
+  class_attribute :solr_fl_parameter, default: 'id,label_ssi,mime_type_ssi'
+  class_attribute :desired_mime_type, default: "application/pdf"
+
+  def perform
+    count = ActiveFedora::SolrService.count(solr_q_parameter)
+
+    (0..(1 + (count / solr_page_size))).each do |page|
+      ActiveFedora::SolrService.query(solr_q_parameter,
+                                      fl: solr_fl_parameter,
+                                      rows: solr_page_size,
+                                      start: page * solr_page_size).each do |document|
+        if document[:mime_type_ssi] == desired_mime_type
+          # Given that we have a mime_type we can assume that we've successfully attached the file.
+          ConditionallyResplitFileSetJob.perform_later(file_set_id: document[:id])
+        else
+          # We have failed to attach the file to the work.
+          ConditionallyReingestFileSetJob.perform_later(file_set_id: document[:id])
+        end
+      end
+    end
+  end
+
+  ##
+  # A helper module for conditionally finding a file set.
+  #
+  # @see #find
+  module FileSetFinder
+    ##
+    # @param file_set_id [String]
+    # @return [FileSet] when the given :file_set_id is found.
+    # @return [FalseClass] when the given :file_set_id is not found.
+    def self.find(file_set_id:)
+      FileSet.find(file_set_id)
+    rescue ActiveFedora::ObjectNotFoundError
+      message = "#{self.class}##{__method__} unable to find FileSet with ID=#{file_set_id}.  " \
+                "It may have been deleted between the enqueuing of this job and running this job."
+      Rails.logger.warning(message)
+      return false
+    end
+  end
+
+  ##
+  # This job conditionally re-splits a file_set's PDF.  How do we know if we need to re-split
+  # it?  See the {#perform} method for details.
+  #
+  # 1. The file_set is a PDF.
+  # 2. The file_set's PDF is one that we would normally split.
+  # 3. The file_set's parent does not have child works; the assumption being that if it doesn't
+  #    have child works, then
+  class ConditionallyResplitFileSetJob < ApplicationJob
+    ##
+    # @param file_set_id [String]
+    #
+    # @return [Symbol] A terse explanation of what was done with this job.
+    #
+    # @raise [ActiveFedora::ObjectNotFoundError] when the given FileSet's parent could not be found.
+    # rubocop:disable Metrics/LineLength
+    def perform(file_set_id:)
+      file_set = FileSetFinder.find(file_set_id: file_set_id)
+
+      # We've logged this (see FileSetFinder.find) so we'll move along.
+      return :file_set_not_found unless file_set
+
+      # When we aren't working with a PDF, let's not proceed.
+      return :not_a_pdf unless file_set.pdf?
+
+      # When the PDF we are working with isn't something we split, let's bail.
+      return :non_splitting_pdf unless IiifPrint::SplitPdfs::AdventistPagesToJpgsSplitter.split_for?(path: file_set.label)
+
+      parent = IiifPrint.parent_for(file_set)
+
+      raise ActiveFedora::ObjectNotFoundError, "Expected #{file_set.class} ID=#{file_set.id} to have a parent record." unless parent
+
+      return :parent_does_not_split unless parent.try(:iiif_print_config).try(:pdf_splitter_service)
+
+      # When the parent has children, assume that we've already previously succeeded on splitting
+      # this PDF.
+      return :has_children if parent.child_work_ids.any?
+
+      IiifPrint::Jobs::RequestSplitPdfJob.perform_later(file_set: file_set, user: User.batch_user)
+      :requesting_split
+    end
+    # rubocop:enable Metrics/LineLength
+  end
+
+  ##
+  #
+  class ConditionallyReingestFileSetJob < ApplicationJob
+    ##
+    # @param file_set_id [String]
+    # @return [Symbol] A terse explanation of what was done with this job.
+    def perform(file_set_id:)
+      file_set = FileSetFinder.find(file_set_id: file_set_id)
+
+      # We've logged this (see FileSetFinder.find) so we'll move along.
+      return :file_set_not_found unless file_set
+
+      # TODO: The file set does not appear to have a properly attached file.
+    end
+  end
+end
diff --git a/lib/iiif_print/split_pdfs/adventist_pages_to_jpgs_splitter.rb b/lib/iiif_print/split_pdfs/adventist_pages_to_jpgs_splitter.rb
index f9fc8700..1b5115e1 100644
--- a/lib/iiif_print/split_pdfs/adventist_pages_to_jpgs_splitter.rb
+++ b/lib/iiif_print/split_pdfs/adventist_pages_to_jpgs_splitter.rb
@@ -3,13 +3,24 @@
 module IiifPrint
   module SplitPdfs
     module AdventistPagesToJpgsSplitter
+      ##
+      # @param path [String] the path, in particular filename (that hopefully ends with an
+      #        extension).
+      #
+      # @param suffixes [Array<String>] the list of suffixes that we want to ignore for splitting.
+      # @return [TrueClass] when we should be splitting this path.
+      # @return [TrueClass] when we should not be splitting this path.
+      def self.split_this?(path:, suffixes: CreateDerivativesJobDecorator::NON_ARCHIVAL_PDF_SUFFIXES)
+        suffixes.none? { |suffix| path.downcase.end_with?(suffix) }
+      end
+
       ##
       # We do not always want to split a PDF; this provides a decision point.
       #
       # @param path [String] the path of the file we're attempting to run derivatives against.
       # @param args [Array<Object>] pass through args
       # @param splitter [IiifPrint::SplitPdfs::BaseSplitter] (for dependency injection)
-      # @param suffix [String] (for dependency injection)
+      # @param suffixes [String] (for dependency injection)
       #
       # @return [Enumerable] when we are going to skip splitting, return an empty array; otherwise return
       #         an instance of {IiifPrint::SplitPdfs::AdventistPagesToJpgsSplitter}.
@@ -20,7 +31,7 @@ def self.call(path,
                     splitter: DerivativeRodeoSplitter,
                     suffixes: CreateDerivativesJobDecorator::NON_ARCHIVAL_PDF_SUFFIXES,
                     **args)
-        return [] if suffixes.any? { |suffix| path.downcase.end_with?(suffix) }
+        return [] unless AdventistPagesToJpgsSplitter.split_this?(path: path, suffixes: suffixes)
 
         splitter.call(path, **args)
       end
diff --git a/spec/factories/file_sets.rb b/spec/factories/file_sets.rb
index 606c6d2a..fc68671e 100644
--- a/spec/factories/file_sets.rb
+++ b/spec/factories/file_sets.rb
@@ -4,9 +4,21 @@
   factory :file_set do
     transient do
       user { FactoryBot.create(:user) }
+      content { nil }
     end
+
     after(:build) do |fs, evaluator|
       fs.apply_depositor_metadata evaluator.user
     end
+
+    factory :file_with_work do
+      after(:build) do |file, _evaluator|
+        file.title = ['testfile']
+      end
+      after(:create) do |file, evaluator|
+        Hydra::Works::UploadFileToFileSet.call(file, evaluator.content) if evaluator.content
+        create(:generic_work, user: evaluator.user).members << file
+      end
+    end
   end
 end
diff --git a/spec/fixtures/latex.pdf b/spec/fixtures/latex.pdf
new file mode 100644
index 00000000..bdbe71eb
Binary files /dev/null and b/spec/fixtures/latex.pdf differ
diff --git a/spec/iiif_print/split_pdfs/adventist_pages_to_jpgs_splitter_spec.rb b/spec/iiif_print/split_pdfs/adventist_pages_to_jpgs_splitter_spec.rb
index 352b7989..f948eda4 100644
--- a/spec/iiif_print/split_pdfs/adventist_pages_to_jpgs_splitter_spec.rb
+++ b/spec/iiif_print/split_pdfs/adventist_pages_to_jpgs_splitter_spec.rb
@@ -3,6 +3,23 @@
 require 'spec_helper'
 
 RSpec.describe IiifPrint::SplitPdfs::AdventistPagesToJpgsSplitter do
+  describe '.split_this?' do
+    subject { described_class.split_this?(path: path) }
+
+    [
+      ["hello.jpg", true],
+      ["hello.reader.pdf", false],
+      ["hello.reader.jpg", true],
+      ["hello.reader.pdf.pdf", true]
+    ].each do |given_path, expected_value|
+      context "given #{given_path.inspect}" do
+        let(:path) { given_path }
+
+        it { is_expected.to eq(expected_value) }
+      end
+    end
+  end
+
   describe '.call' do
     subject { described_class.call(path, suffixes: ["spec.rb"], file_set: create(:file_set)) }
 
diff --git a/spec/jobs/file_sets_reprocess_job_spec.rb b/spec/jobs/file_sets_reprocess_job_spec.rb
new file mode 100644
index 00000000..b40a76e5
--- /dev/null
+++ b/spec/jobs/file_sets_reprocess_job_spec.rb
@@ -0,0 +1,19 @@
+# frozen_string_literal: true
+
+require 'spec_helper'
+
+RSpec.describe FileSetsReprocessJob, clean: true do
+  describe '#perform' do
+    let(:user) { FactoryBot.create(:user) }
+    let(:file_set) { FactoryBot.create(:file_with_work, content: file_content, user: user) }
+    let(:file_content) { File.open(fixture_path + '/latex.pdf') }
+
+    it 'submits jobs' do
+      expect(described_class::ConditionallyResplitFileSetJob).to receive(:perform_later).with(file_set_id: file_set.id)
+      file_set
+
+      described_class.perform_now
+      # Verifying that we found one record to consider resplitting.
+    end
+  end
+end