Could changing the criteria based on demand lead to a perception of unfairness? Is it right to say that, if I brought you a case now that was a score four you would look at it, but if I brought you the same case in six months’ time and you did not have the extra resource, then you would not?

That is absolutely right, and we have recently identified this as a concern. As a result, under the public value model we are going to change the wording that we use to notify complainants of a decision. We will say that we have measured their case at—for instance—a score three, but that they should be aware this is subject to change depending on the pressure on our services and difficult decisions that we might have to make.

You have talked about how you communicate with complainants. When a caseworker decides not to take a complaint forward, how much detail do they give the complainants as to how they arrived at the scoring?

Complainants are given information about how the case was decided, but we want to improve this going forward. It is very important to us that complainants are happy with the decision that is made, and our experience is that when that conversation is had properly, people understand the reasons. What we need to do better is to manage their expectations. That is why, from next month, we will be changing our website to give a clearer explanation of how the public value model operates. We did not do this when we first launched it a year ago because we were concerned about getting the balance of information right. For instance, we know that AI is deployed by complainants and we wanted to make sure that we got a fair result. We came to the conclusion that we want to own this. We recognise that AI is out there, and we will have prompts on our website for complainants that will say, “If you’re going to be using AI, these are the kind of questions you might want to ask it. This is how we want it to be operationalised.” Initially we did not do that because we wanted to wait until we changed the score in November, and then in February, as I said, we had another conversation where we went back and made different changes; but we now think we need to make that information available. It is in line with our transparency value, and we will be doing that in April.

Going back to an earlier question, are you worried about the risk that people may self-score? For instance, they might say, “Oh, it’s only a two, I won’t send it in.” If that happens, you lose your data collection.

That is a worry, but as Paula said earlier, at the moment it has not manifested itself in terms of complaints coming to us. The rate of demand is only going one way, but we need to get that communication right so that people understand that we will use their data whether or not we investigate their case. We do not have the capability yet, but we hope to develop a mechanism whereby we will be able to come back to complainants subsequently and say, “You complained to us two years ago. We didn’t investigate your complaint because it fell under our threshold, but we did take your data; there is a systemic issue that we later identified, so we carried out this investigation and this was the impact and the result.” That would be a great thing to be able to do, but we are not there yet.

So they would get feedback that they had influenced a positive change to the system?

Yes.

When you talked about the scoring, I noticed that you started with two things that are very much about the complainant—the impact it has on them, and whether it is influenced by any protected characteristics—and yet you call it a public value model. Do you think that is a good name, or do you think it might imply that a case is not publicly important enough to follow?

We have always called it the public value model. It is about the impact on the way systems are developed: as Paula was saying earlier, there has to be a balance between individual cases and systemic impact. Our job over the strategic period is to make sure we have that balance. By way of an example, one of the access to justice criteria on which we score complaints is whether a disabled person has difficulty accessing our systems. Under the public value model, we will give that complaint a priority so that we can give them the support they need; but the public good is also involved because of course we want public services to be available to people who are disabled and need additional support. So all these things are slightly nuanced, both in terms of the individual case and the greater systemic impact.

You have been very clear that demand for your service is continuing to rise; there are various reasons for that, and research is ongoing. Is there any sense that this could be due to a rise in the underlying problem of maladministration, or is that not the case?

We suspect that the way in which public service is being delivered accounts for that rise. Rigorous research on the contributing factors is under way now, and we are keen to come back to this Committee once we have the results. Our hypothesis centres on the quality of public service, levels of trust in it, and the fact that artificial intelligence has accelerated the mechanism for raising concern. But I would rather the answers came from the research. Importantly, there may be some outlying factors that will potentially be a feature in our demand going forward.

I look forward to seeing that. If it is indeed the case that public services have seen more maladministration and more underlying problems despite all the recommendations you are giving them to improve, why is that happening? You can decline to answer if you think it is best to wait until after the research concludes.

I would rather wait until the research comes back, but I would say that 80-85% of our work is in the health sector, so delays and waiting times are always a big factor. That goes back to rising demand and the way in which resources are deployed. We will look for impacts on the individual and across systems, which we are doing now and will do assiduously throughout our strategy, but there may be factors that are outwith our influence.

We have just been talking about the public value model. How are you measuring the impact of that model on your service delivery and outcomes?

Rebecca has taken us through the issue of demand. In terms of impact, for me it is about the outcomes, both for the individual and for organisations. Some recommendations can be small but meaningful for the local community. We are increasingly looking at things that are structural and systemic, particularly in the health service, such as changes in policy, changes in structure, or changes in practice that we have influenced or driven. I would always stress that we achieve this through our voice: we are extremely keen to do more work with Parliament, and in partnership with key bodies that have a different set of powers and different jurisdictions.

I would like to move on to a couple of questions that have come up about AI. First, the use of the algorithmic transparency recording standard has recently been made mandatory across Government Departments and its arm’s length bodies. Although the PHSO is an independent body accountable to Parliament, I do not suppose it comes under your mandate, but are you able to confirm whether you currently subscribe to the standard or whether you plan to in the future?

Yes, our ethics and transparency policy states that we subscribe to it and we will continue to do so.

Secondly, I refer to a point that Rebecca mentioned earlier about people who are dissatisfied with a complaint outcome and send you 20 pages of details. We get that too, so I can sympathise. How are you dealing with and getting around the impact of that? Are you going through it all, or are you saying, “Look, this clearly has a lot of AI hallucinations, therefore we’re not going to spend much time on it.” How are you sorting through it?

Being the resident tech geek—relatively speaking—I am happy to answer. As Rebecca said, there is easily findable information on our website which says, “If you are using AI, please use it in the following way,” and then describes the prompts. That will probably take us so far, but then we are into the territory of training the models that people are using to use the right data sets and refer to the right legislation. As I am sure you know, that can be done painfully by repeatedly telling ChatGPT, “No, do not go to that statute,” but there are other ways in which we can directly access which sets of data we want these models to refer to. Rebecca has a very strong executive team, including in the technology area, so we are watching this very carefully. But the important thing is, as Rebecca said at the start, that we know AI is a force for the good as well as a force for the bad; like a number of other organisations, we are watching it positively, but carefully.

In the meantime, how do you deal with cases where you get a huge dossier of information around why you need to revisit your decision or whatever it may be? Are you are going through that in great detail?

We still need to go through it. Obviously, we will start to spot patterns: for instance, we will pick up quickly if the jurisdiction is wrong and it is something that the LGSCO should be looking at. But at the moment, we are having to receive it as we would any complaint that was generated in the analogue way.

We are taking on some additional resource in our legal team to deal with this, but we recognise it as being one of the pressures on our resource that impacts everything else, so we are trying to manage it more effectively.

You will be aware that as a Committee, we held a call for evidence and opened a survey for complainants to inform our scrutiny of your work. I caveat that, of course, the survey sample is both small and critically self-selecting. It has to be said that the view of the PHSO that comes across in our evidence is not exactly glowingly positive. What confidence do you have that the new strategy will address this negative perception?

We are always sorry to hear that a complainant is distressed and unhappy. As I said before, I have been spending time with these complainants to understand the drivers, and it is the nature of the beast that we find a high correlation between dissatisfaction and a complaint that has not been fully upheld. Even if we have partly upheld it, the complainant may say, “This didn’t go far enough for me and I’m still aggrieved.” But we take learnings from that. A lot of it relates to delay, so we have talked about how we need to address that. Some is also due to communication with the complainant, and there are a number of ways in which we want to improve that. Some relates to the user experience, and as we say, improving the user experience is part of our strategy. Certainly the PHSO account is going to make it easier for those who can use a digital online account to have a closer relationship with the caseworker, to find out where their case is, and why certain stages might take longer. That will also provide more prompts for the caseworker to follow through and follow up. There are a number of issues about service delivery that we can improve, both through technology and through the training and operational improvements that are coming through the ombudsman assurance team.

How do you help those who are not digitally linked or comfortable online, or who do not have the bandwidth or time to be engaging in that way?

It is important that all existing channels—critically, the phonelines—stay open. I have sat in on a number of calls, and it is a very strong service. Caseworkers speak to the complainant by phone, if the complainant is not comfortable using an account they can continue to use email and so on as they do at the moment. We will keep all those channels open.

Are all those calls handled by a person? If someone is just calling up to ask where they are in the process, can our great friend the bot—or whatever it is called—be used? The bots are so good now, one does not even know when one is talking to them.

If a complainant is happy to use the PHSO account, then potentially down the line they can receive updates through a chatbot, and can ask the chatbot, “Where is my complaint?” If they are not happy to use the account, they can ring up and speak to a person.

Are they able to ring up and think they are talking to a person?

They will know they are talking to a person because we only have people on the phone lines.

Right. What potential is there, without any diminution of service, to deploy technology to answer basic questions from those who do not want to use the online platforms? That would free up more time for actual people to carry out higher level work.

As we see what the uptake of the PHSO account is, there will be some technologies we can deploy through what we call intake. It is all possible, and those are in train.

Can I just add to that? First, our intake workers do a fantastic job: if you ring up intake, you will get through in about three and a half minutes, and you will talk to somebody who has been trained in dealing with people who are distressed and need a whole range of support, empathy, understanding, and signposting. I am not saying a bot could not do that, but our human interface is really excellent and adds a huge amount of value that you will not see in the casework stats because it happens before that. These are people who have over 100,000 points of contact every year. Secondly, I would come back briefly to the point about your survey. We approach this from a learning perspective, and I am really interested in the information you have given. I would like to contextualise it though: we get over 6,000 people completing our surveys every year, so we have a slightly richer data set. As Paula says, for understandable reasons it is very outcome dependent. On average, people tend to be 40%-50% content with our service, but that includes people whose complaints we have upheld, 86% of whom say that we do a good job. Of those whose complaint we did not uphold, 17% disagree. We get a higher approval rating among the organisations we investigate, largely because we deal with the same complaints handling teams, and they are not invested in the outcome in the same way. But we get some really helpful qualitative feedback from them in terms of what they find we do well, and what they would like us to do differently. For example, one concern is about how we deal with differing views about the maladministration, and whether we can do that in a different way. However, the truth of the matter is that the results of these surveys have been the same for years, regardless of various different improvements that we have made, so we do not actually think we are asking the right questions. As of last month therefore we have completely redeveloped the survey to focus on our strategies and ask questions around the three aims that we are trying to achieve. For instance, people will be asked, “What was it like for you to go through the scoping exercise with your caseworker?” Or, “What was it like for you to look at the preliminary provisional views?” It will be almost like a pulse survey in relation to a particular part of the process, which of course is completely disengaged from the outcome, and it will therefore be more balanced and objective. We will be using that survey at the close of each case, whether it has been through early consideration or primary or detailed investigation, using the portals that Paula has been talking about. Around 600 people have already used the survey and we will be reporting on it from next quarter.

Just on that last point—while accepting that there is likely to be a sampling issue as people will be more inclined to respond if they are upset—in the survey the overwhelming majority of respondents waited very long periods of time for their cases to be concluded. I am wondering if the perception difference might have an impact on this. From a local authority background, we quite often have a case that comes in and gets shut down; then a separate case for that same person comes up with the same issues. From their perspective it is a single case, but from the organisation’s perspective there are multiple cases. Is there anything in the way in which you process cases that could explain that disconnect?

It is difficult to tell, to be honest, without knowing which cases these were; as you say, it is only a small number. However, we could be more aware that sometimes there is a disconnection between the different stages of a complaint. I will give you an example that I came across recently. As Paula says, we get complaints directly to our inboxes, so we have quite a good understanding of how things can go wrong. This individual had made a complaint—I cannot remember which body it was about—but it transpired that this body still wanted to continue some first instance decision-making. So after reaching the end of the queue, the case was closed pending further complaint handling by the organisation concerned. Once that stage was completed and the individual came back to us, they were put at the back of the queue. There is a reason for doing that, but it is not very helpful and we need to be smart about how we deal with it. We need to ensure that we are always looking at our service from a user perspective; that is one of our key aims, and it will guide us in our new strategy.

We could do more to convey how the process runs, what takes time, and what is going on behind the scenes. I went down to an MP’s office to see how his constituency operates in taking complaints and referring them to us. He is a very established MP and has been running his constituency office for a long time; he made the interesting observation that when his constituents say, “Why is this taking so long?” He replies, “It’s because you want this process to be meticulous, and meticulous takes time.” Sometimes it absolutely does, but we could be clearer about the depth of work we have to do to make sure that our analysis is right, particularly in health where we are working with a trust to make those changes. We should be looking at the time taken from the user’s perspective as well, standing in their shoes and keeping them up to speed and up to date. I am sorry to harp on about the portal, but it is important that we get this account absolutely right. There is a particular complainant who I have been talking to over the last few months. She had an acute experience with both the trust and with us; happily, she has a systems and technology background, so she has given us input and has been going through what she was dealing with as we design every factor of the portal, so that is going to be particularly helpful. We need to get this as right as we can.

Peter, you said you came from a local authority background; one thing that the Local Government and Social Care Ombudsman does well is group cases. It is able to do better than us because it has a different constituency in terms of the organisations it investigates. If there is a particular local authority and the LGSCO sees lots of complaints about the same service in that local authority, it is able to look at them all together. That is something we want to do more because we think it will help us to manage waiting times. But if you have a local authority background, you have probably seen that going on.

I am really just inquiring as to whether there might be a methodological explanation for the difference in terms of timings, but I fully accept the point around meticulousness. Turning to my actual question, during the course of our inquiry, issues with interpretation and communication of the 12-month time limit for accepting the complaint have been highlighted as a particular problem. Are you looking at ways in which you can ensure the guidance could be applied to be more consistent?

If I could just start: the 12-month time limit is statutory, so there is nothing we can do about it, short of primary legislation. We cannot take cases that are more than 12 months old, subject to using our discretion, which has to be done in a reasonable way. To be honest, where there is a really good reason for being late, and where we have seen that the complainant has been making consistent efforts to go through the earlier stages, then we will be sensible and proportionate in terms of waiving that limit. I have personally worked on cases where that has happened. But on the other hand we have to recognise what happens if we do not do that because we are then investigating historical events where the evidence is harder to access, people have moved on, and their memories have gone. Often events have been overtaken and the issue has been resolved. In the meantime, we are not investigating more pressing, more recent issues because our resources are being sucked up into looking at something that is several years old. So there is a very good reason for that time limit, and in terms of using discretion, our caseworkers look at it from a perspective of what is sensible.

There is so much, if I can build on that—

I am conscious of time, so can I implore focus?

Exploring that further, in terms of the guidance itself, do you set out the reasons why the 12-month time limit makes sense—as you have just done here—and make the circumstances in which discretion may be exercised very clear?

Yes.

You do; great.

That was nice and short.

In terms of that time limit, you said that where someone is making consistent efforts to make a difference, very often you will encourage people to try to make sure they use the full process available before coming to the PHSO in order to reduce the burden on you. However, if 12 months really is the timeline, maybe we need to be advertising that option earlier. Very often, the people we are dealing with tend to be higher agency people who know how to access us; people with lower agency need to have a greater awareness of what their rights are and how to access the service. What efforts have been made to ensure that is the case?

If I can quickly come in on that; we work closely with 650 MPs and communications are really important. You are incredibly important in helping us process complaints, and we are delighted therefore that the January survey said we are raising awareness with you. We must make sure we communicate clearly to you so that you can clearly communicate to your constituents; that is absolutely part of it.

It is maybe a public awareness campaign as well.

Yes, we will use all those channels.

When we came up for our visit last year, which I found very interesting, you mentioned the MP filter and the potential issues with it. A respondent to our inquiry described it, in their exact words, as, “A gatekeeping mechanism that denies access to administrative justice.” Would you agree with that?

I probably would not describe it in exactly those same words. The evidence indicates that it does deter complainants from making complaints about government services. On balance, yes, we would still want it removed, but as I just described to your colleague, we think that working more closely with constituency MPs and select committees, and raising awareness, will help increase that complaint flow about government services.

Probably most of us would agree with removing the necessity to go through an MP, but do you envisage any unintended consequences of doing that? Do you think your workload would increase beyond what you currently see?

Yes, it could increase but that is where the public value model, which is about being able to calibrate and manage that, comes into its own. We anticipate that we would be able to manage that increase were the legislation to be changed.

The Government have now retaken their decision in relation to the PHSO’s investigation into communications around changes to the women’s state pension age, and will not be providing compensation. There is obviously a range of views about that decision, even on this Committee. I personally find it deeply disappointing and regrettable. The DWP acknowledges that maladministration happened, but Paula, you in particular have expressed what I would describe—in my words, not yours—a level of frustration with the DWP around communicating action plans from the lessons learned. There has also been some mischaracterisation of the findings about the potential effect of the letters that should have been sent earlier. How does the way the DWP has been engaging with the PHSO on this issue compare with other organisations that you investigate, particularly when you may make a recommendation that is not followed up?

There is quite a wide spectrum of engagement and responses. Operational relationships with the DWP are very good. Rebecca referred to the sessions that she and I have been running for senior operational colleagues, which are very citizen-centric: they have demonstrated much willingness and alertness to the issues we raise. In the case of the pension issue, it is now 24 months since the inception of the action plan and it is disappointing that we are not further forward. We are working as constructively as we can with the Department, but we will need sustained senior engagement. That is a necessary ingredient in getting the action plan through.

Do you have a timetable for when the action plan will be produced?

We would like to start work on it as soon as possible.

You have not started work on it yet?

Not yet, no.

Are you confident that the DWP can change its culture to one of continuous improvement, openness and learning from feedback, as you were speaking about earlier in the session, when engaging with trusts, for example?

We already see a lot of that. It was extremely useful to have had these sessions to review structures and mechanisms, opportunities and challenges. In certain parts of the organisation we see a huge focus on this, and it absolutely gives us reason for confidence.

There is no enforceable mechanism for Parliament to ensure that your recommendations are adhered to, and there are obviously limits to public scrutiny. Are you concerned about the impact that the Government’s decision on the WASPI women will have on public confidence in the PHSO? Will you be advocating for any changes in powers and responsibilities in light of that?

We would rather not advocate for powers because, as I said right at the top of the session, our aim is to work constructively with public services. The decision is disappointing but I do not think it is widely symptomatic. It is quite an exceptional case, but we are very grateful for all the parliamentary committees that will work with us when we want to highlight action. We are very grateful to this Committee and the Work and Pension Committee for their engagement on this topic, and we think we can do more in that regard.

Can I just add to that? Our compliance rate across all our cases is over 99%, and these conversations with Government and the DWP are outliers. Our impact is not just made through casework, although that is obviously very important; we also have impact through our insight reports. For example, our report a couple of years ago on discharge from mental health inpatient care had a real impact on recommendations picked up in the Mental Health Act. I will keep my comments short, but there are many other examples. As Paula just mentioned, going forward we would like to build on our work through parliamentary committees and in particular with PACAC. We are not worried about what that means in terms of our reputational impact because a key aim of our new strategy is about building that voice and using our systemic findings to have greater impact nationally in terms of public services. We are very confident about that.

Paula, you said in our pre-appointment hearing that you wanted to look at governance. Have you had the opportunity to make an assessment of the governance arrangements at the PHSO so far?

Yes, fairly extensively. Turning first to the board and its committees, we do an effectiveness review every few years, which has quite a strong score. It is a strong board with two subcommittees that really focus on the assurance side. But we have just put in place a slightly different model for the board that leans more towards strategy and influence, including bringing in some board members who can really support us with that, both in health and in Government. Secondly, we are also looking at our end-to-end assurance; increasingly, in working with what we call the public engagement and advisory group, which is a great group of complainants and non-complainants. We are seeing how its voice can be brought forward into our board meetings. So some progress has been made and we are happy with it, but we can go further.

You also decided to abolish the quality committee and the inclusion and wellbeing committee, and to set up a strategic delivery committee. Is that largely focusing on casework assurance and the experiences of service users, or does it also combine with corporate strategy?

The work of those committees has been absorbed either within the board or the strategic delivery committee. If you take the assessment of wellbeing and EDI, that comes up directly into the board. The strategic delivery committee takes the other aspects into account, which includes all our planning. Audit and risk continues to do risk assurance. In that risk register, we look at the inherent risk on our caseload as well. So I would describe it as having been reconfigured.

I would just add that those changes were made before Paula’s appointment. The reason why the quality committee was set up in the first place was to oversee a quality programme that was brought in many years ago to support a necessary improvement in quality output from casework. That programme came to an end a couple of years ago, and that was the reason why the quality committee was closed down. Paula has spoken about the other pieces there.

Is the board developing a new equality, diversity, and inclusion strategy as part of its work?

EDI comes through our people reporting and board members and, as chair of the board, I will scrutinise the executive on where we are on that. It is a regular item coming through the board meetings.

So you are not in the process of doing a new strategy; it is about scrutinising an existing strategy in that space?

Correct.

Do you know when that is due for a refresh?

That is a very good question; I do not, but I am happy to come back on that point.

Thank you, colleagues. Do you have any further questions for our two witnesses this morning? No? In that case, thank you both very much indeed for your attendance and for taking our questions.