Tools and providers
Part of Procuring corporate communications tools with a dossier before a shortlist
Selecting corporate communications software only after the manual route is costed
Admit and select communication software for one England job using exact service evidence, a manual baseline and independent rejection gates.
Selection should begin with an admission rule. For fictional Cedar Vale Engineering Ltd, the job is to control one source copy about a proposed England customer-support site consolidation, route it through authorised reviewers and preserve correction evidence. The buyer-owned manual process is the first candidate.
Cost the manual baseline before software research. For Cedar Vale, an illustrative worked example is 6 staff hours per approved source copy: 2 hours drafting, 2 hours reviewer checks, 1 hour accessible-format preparation and 1 hour correction or version recovery.
At an illustrative loaded rate of GBP 45 per hour, that is GBP 270 per copy. A late correction reaching 400 employees can add 3 hours of reissue work, so an illustrative repeat cycle costs about GBP 405. These figures are illustrative only and depend on the buyer's own volumes, roles and pay rates.
No hands-on trial was completed. The providers named below are examples of what exists in each function, not tested or ranked products, and the best tools overview covers wider options. This method uses evidence checked on 6 September 2026.
What to take away
- Start selection by costing the manual process before researching any corporate communications software product.
- Keep an evidence index for supplier claims, contract terms and expiry triggers.
- Test acceptance with synthetic data and disabled production publishing before comparing candidates with the manual baseline.
Describe the missing capability
Map the current steps before writing requirements. State the single failure the buyer wants to remove. Leave volumes, response times, prices and weights blank until evidence exists.
Map current communication steps
- Source copy created
- Facts approved
- Employee boundaries handled
- Accessible formats prepared
- Publication authorised
- Objections received
- Obsolete version corrected
Do not buy one vague communications platform. Decide whether the gap concerns newsroom publishing, employee messages, contact email, monitoring, social publishing, research, approval, incident work or analytics. A tool admitted for one function receives no credit for another without separate evidence.
Name candidates for one function only
Once the single function is chosen, list candidates for that function alone. These established providers are starting points, not endorsements:
- Newsroom and press release distribution
- Cision, Prowly.
- Media monitoring
- Meltwater, Cision.
- Employee communications
- Staffbase, Poppulo, Workvivo.
- Social publishing
- Hootsuite, Sprout Social.
- Contact email
- Mailchimp.
Each name still has to pass the admission test below. None has been tested here, and none is admitted for a function outside its row.
Apply the document admission test
Require the legal supplier entity, trading name, exact service and edition, version, UK availability, terms, price basis, support scope and change policy. The Companies House register search can identify the legal entity and its filed details. Registration does not prove that a service exists or suits the buyer.
Reject a candidate when a material field is missing. "Contact sales" is not a price; "enterprise grade" is not security evidence; a general accessibility statement does not show the configured task works.
Worked example: a candidate has no price basis and no named support scope. Those are two hard-gate failures, so the candidate is rejected even if its social publishing screen looks useful.
Create an evidence index, not a scorecard of pasted claims. For each document, record URL or file identifier, publisher, service edition, publication or update date, access date. Add territory, customer type, claim supported and expiry trigger.
Supplier sales material remains first-party evidence. A contract term may supersede a public description for the buyer, but the reviewer must record the conflict rather than silently selecting the more convenient version.
Inspect the data arrangement
Draw the path for employee, customer, supplier and media-contact information separately. Record who determines purposes and essential means, instructions, subprocessors, international access, retention, rights support and deletion. The ICO's controller and processor contract guidance is under review following DUAA and describes Article 28 arrangements. Legal labels still need to match the facts.
For a possible restricted transfer, use the ICO's brief transfer guide and check its current version. Supplier location, server location and recipient legal entity should not be guessed from branding.
Set observable acceptance work
Use synthetic identities and a clearly fictional source document. Disable production publishing and delivery, as the guide to implementing a tool with production switched off describes. Test access roles, approval order, version recovery, correction, audit export, accessible use, account recovery and deletion. The NCSC's cloud security principles offer buyer questions about identity, administration, audit and secure use; they do not certify a product.
Each observation gets an evidence ID, configuration, tester, time and limitation. A pass means only that the tested event occurred. Unavailable evidence is unresolved, not assumed.
Compare the same unit at decision time: one approved sample moving from author to staged correction and evidence export. Do not award points for extra modules Cedar Vale has not admitted. The procurement owner should also document whether the manual route can be repaired more safely than a new system can be introduced.
The procurement owner compares the admitted candidate with the manual baseline. Legal, privacy, security, accessibility, commercial, finance and exit gates are independent. Any failed hard gate ends selection, even if another function looks convenient. Repeat the supplier-record check on the day of purchase.
Compare a candidate with the manual baseline
Compare cost per approved source copy, not licence price alone, using the supplier comparison method. Take the illustrative baseline of GBP 270 per copy and assume 40 copies a year, which is GBP 10,800 of manual time. Suppose a candidate removes version recovery and accessible-format preparation, saving 2 hours or GBP 90 per copy, so GBP 3,600 a year.
If that candidate's illustrative annual price is GBP 6,000, it costs GBP 2,400 more than the manual route and fails the cost gate. At GBP 2,000 it would pass the cost gate, but it would still face the legal, privacy, security, accessibility and exit gates. Under these illustrative numbers, the manual route stays.
Before you act
- Map current steps and record the actual failure to remove.
- Decide which single function the gap concerns.
- Require legal entity, edition, version, price basis and support scope.
- Build an evidence index with dates, territory and expiry triggers.
- Draw data paths for each contact group separately.
- Run acceptance work with synthetic identities and publishing disabled.
Common questions
What should a buyer do before writing requirements?
Map the current steps first. Note who creates the source copy, approves facts, handles employee boundaries, prepares accessible formats, authorises publication, receives objections and corrects an obsolete version. Record the actual failure the buyer wants to remove, and leave volumes, response times, prices and weights blank.
How should a buyer treat supplier sales material as evidence?
Supplier sales material remains first-party evidence. For every document, capture the URL or file identifier, publisher, service edition and dates. Also record territory, customer type, claim supported and expiry trigger.
What does a pass in acceptance testing actually prove?
Acceptance testing proves only that the tested scenario worked in the stated configuration. It does not prove that untested roles, integrations, accessibility settings or future changes will work. Nor does it replace unresolved supplier evidence or the buyer's own risk decision.
