My Voice, Better Supported: AI and Speech Pathology

A conversation with Isabel Lund on 9 November 2026, and a practical account of how speech pathology, AI assistance and my own voice can work together.

I took the speech pathology call from my car. Around me were the ordinary complications of a mobile life: a phone, an iPad, charging equipment, notes to keep within reach and a head full of things I wanted to explain before I forgot them.

I wanted to talk about technology. I wanted to talk about writing. I wanted someone to understand why an email can take several attempts, why a spoken question sometimes disappears while I am constructing the answer, and why the polished paragraph that eventually reaches an agency can conceal the effort behind it.

Isabel gave me time to explain. She also kept bringing us back to the assessment. I appreciated both. By the end, we were discussing something much larger than a device or subscription: how I could communicate with more confidence, participate in the community and keep my intentions intact as my words travelled from thought to speech to paper.

This is a thank-you to Isabel and the team at Communicate Speech Pathology. It is also a participant's account of what AI-assisted communication looks like in practice.

I came for speech pathology as well as technology

It would be easy to describe the appointment as a request for AI funding. That would leave out one of the clearest things I said.

“Speech pathology is important, irrespective of what I'm going to get in the next plan or assistive tech.”

My words as captured in source excerpt 2.

I spoke about attending speech pathology as a child for stuttering. I described a continuing difficulty that I call a “mental stutter”: the gap between what I intend to say and what comes out. Sometimes I choose the wrong word. Sometimes I interrupt because the thought is urgent. Sometimes I am already following a connection while the person speaking is still finishing the question.

Those descriptions are my experience. Isabel's work is to assess what is happening and which approaches may help. The appointment gave us room to explore speech, comprehension, memory demands, social communication, written expression and the situations in which I need support.

What I valued was her willingness to look at the whole communication task. I can have a substantial vocabulary, tell a story and care deeply about the subject, while still finding a particular exchange difficult. In this conversation, those strengths and difficulties could sit alongside one another.

The small moments that explain the bigger problem

A long report can contain a thousand claims about functional capacity. Sometimes the most revealing evidence is a short question.

“What was the question again?”

My request for repetition appears in source excerpt 40.

During an everyday planning task, I had to hold several errands and closing times in mind, then explain an order for completing them. I asked for the information again. Isabel repeated it and gave me another opportunity to reason through it.

“Yeah, absolutely. We can have another go.”

Isabel's response as captured in source excerpt 41.

I remember the practical kindness of that exchange. I was trying to follow the task, asking questions and explaining where I lost the thread. Repetition gave me another way into it.

There was another useful moment when I gave a phone number quickly in the grouping I use to remember it. The exchange led to a discussion about slowing down and breaking numbers into manageable parts for the listener. I recognised the problem: the arrangement that helps me recall something may make it harder for another person to receive it.

Isabel also recognised strategies I was already using, including describing a word and using a first-letter cue. The captured text includes:

“So you're already using some good strategies.”

Isabel's words in source excerpt 44.

That encouragement mattered. I want to build on what works. I want practice, feedback and ways to make the next interaction easier. These examples show the work we discussed. Any interpretation of my assessment performance, including scores or diagnostic conclusions, belongs in Isabel's clinical record.

Why a fluent email can hide a difficult process

My writing can look confident. The process behind it is often a sequence of speaking, stopping, dictating, listening, correcting, finding an earlier document and trying again.

During the assessment I described difficulty reading bureaucratic forms and understanding what the writer is asking me to do. I explained that I put material through AI to help me work through it. I also talked about listening to audiobooks and replaying sections. With my reported vision difficulties, audio is an important part of how I access information.

In practical terms, I use AI to help bring a long passage into a form I can work with. I can ask for the questions in order, a plain-language explanation, a draft response or a list of points I still need to check. I can then listen, compare the output with the source and correct it.

The final email is only the visible end of that process. The preparation may have helped me retain the question, organise the answer and return to the point after a distraction. When I describe AI as a cognitive aid, this is what I mean.

It also has limitations I need to manage. A fluent sentence can contain an inaccurate date, a misheard name or a stronger claim than the source supports. My correspondence needs checking precisely because it matters to me.

Isabel understood that communication includes the reader

I love detail. I love a thread that leads to another document, another date, another explanation. That instinct helps me investigate. It can also leave a reader struggling to find the question I need answered.

The assessment explored how much information I give, how I start a conversation, how I respond to disagreement and what happens when I feel misunderstood. I acknowledged that I can interrupt, become blunt or send repeated emails. I also said I am working on those difficulties.

That is an important part of my account. I want my concerns understood, and I want better ways to express them. I want to notice when I have lost the listener, clarify what I mean and repair a conversation before frustration takes over.

Isabel introduced the phrase “double empathy problem” during our discussion of misunderstandings between autistic and non-autistic people. I took it as an invitation to think about how understanding can break down between people with different communication experiences. We can work on my skills and also ask communication partners to allow processing time, clarify their meaning and check what each person understood.

Her approach gave me something constructive to work towards. Being understood involves the message, its timing, the amount of detail and the person receiving it.

A prompt that brings me back to the purpose

One of Isabel's practical suggestions concerned prompts. If I went to a doctor with two things to discuss, having those points available could help me return to the purpose of the appointment when my thoughts moved elsewhere.

I recognised that immediately. I had prepared notes for this very call.

“That's exactly what I do.”

My response as captured in source excerpt 76.

This is an area where my existing technology use and speech pathology could fit together well. A clinician can help me choose an effective strategy. I can use an accessible note, a spoken reminder or a short script to carry it into an appointment. Afterwards, we can look at whether it helped.

For me, the format needs attention too. A page of small text may be difficult to use. A brief prompt that I can enlarge or hear may be more practical. The right arrangement is something to assess and practise in the circumstances where I actually need it.

I would like to become better at arriving with a purpose, following the conversation and leaving with a clear understanding of what happens next.

What we discussed about AI and assistive technology

I explained how heavily I already rely on AI for writing, listening, organising material and preparing correspondence. I asked for an honest account of those functional uses in the report.

Isabel treated this as a developing area worth exploring. In the captured exchange, she said she would ask her supervisor for thoughts about advocating for these tools.

“It's a tricky area, I guess, because it is new, but it sounds like it is supporting you a lot.”

Isabel's words as captured in source excerpt 80.

I appreciated that response. She listened to the use I described, acknowledged the support it appeared to provide and proposed getting further advice. That gives us a basis for a more specific discussion about suitability and evidence.

We also discussed an earlier communication-technology proposal, word prediction and how different approaches might fit my way of communicating. Some overlapping transcript fragments conflict on exactly what had previously been trialled. I am leaving those details for confirmation. My aim is to identify the features that help me with real tasks, then work with the relevant professionals to assess them.

OpenAI/ChatGPT and an option I called “Dots” came up in my account. The exact product I meant by “Dots” still needs clarification. The prices and subscription descriptions I mentioned during the call also need checking before they become part of a funding proposal.

At this stage, the conversation establishes a request for assessment and further consideration. A particular AI subscription, device or funding allocation has yet to be settled.

The agent's job in my communication workflow

One of the most practical things I raised was reducing repeated correspondence. When I feel the point has been missed, I can write again, add another explanation and then send another version. I want help keeping the record organised and making the next message easier to follow.

In source excerpt 68, I described wanting agent assistance to avoid “duplicates, triplicates.” That is a concrete aim we can examine.

My intended workflow is straightforward. I supply the account and decide what I want to communicate. The AI helps assemble the material, identify the main request and prepare a readable draft. I review important details and authorise the action. The agent can then carry out the approved administrative steps, including sending a prepared message and attaching the relevant files.

The usefulness depends on how well that process works. Does the message retain my meaning? Are the attachments the right ones? Has an earlier message already covered the same point? Can I find the sent version later? Those are the questions I want the workflow to answer.

The professional relationship remains central. Isabel can assess my communication, help me practise strategies and decide what clinical recommendations she can support. Technology can help me prepare, retain and follow through on that work between appointments.

We used that workflow after the assessment

Towards the end of the call, I offered to send Isabel my notes and a summary of what we had discussed. The captured exchange records her welcoming that offer. I then instructed my agent to prepare the material and send the follow-up.

The outgoing email is recorded in my Sent folder at 12:56 pm AEDT on 9 November 2026. Its subject was “Thank you — 9 November speech pathology assessment notes and follow-up.” It was addressed to Isabel and included two attachments:

  • Assessment_notes_for_Isabel_9_November_2026.txt, a thematic account of the assessment, reconstructed exchanges, questions and proposed follow-up.
  • Assessment_source_excerpts_9_November_2026.txt, the source conversation excerpts retained for cross-checking.

The message thanked Isabel for her time, patience and practical suggestions. It also asked her to correct anything that differed from her own record.

“Please correct anything that does not match your record.”

From the sent follow-up email.

It explained that the notes had been prepared with AI assistance and would support her note-taking and report preparation. It identified the reconstructed passages and the unverified automatic transcript. It asked about functional communication, processing and recall, social participation, self-advocacy, possible assistive options and the proposed fortnightly therapy.

This gives me a specific example of the workflow in action: conversation material became an organised participant submission, with the source text available alongside it, and the approved package was sent to the clinician. The Sent record confirms the outgoing message and attachments. Isabel's receipt, corrections and clinical response are subsequent steps.

Even the transcription errors teach us something

The source pack contains overlapping versions and obvious recognition problems. “Racewalking” becomes “race loving” in one fragment. Product names vary. Some short replies are difficult to assign to a speaker with confidence.

Those errors are a practical reminder of the correction work that comes with transcription. An automatic output can help me retain a conversation, while still requiring comparison with the original recording and the clinician's notes.

The attachments preserve that distinction. The source excerpts are copies of text supplied during the conversation. The thematic notes reconstruct likely speakers and meaning. The short automatic audio transcript has its own limitations. Keeping those forms identifiable allows a disputed passage to be checked.

For the public article, I have selected relevant extracts and left out private contact details, meeting-access information and unrelated personal history. The fuller source pack was sent directly to Isabel for clinical review.

Fortnightly therapy, confidence and a life beyond correspondence

One of the most encouraging parts of the assessment was the discussion of ongoing therapy. Isabel raised advocating for fortnightly speech sessions. I said I wanted that.

We discussed written and spoken communication, social confidence and engagement with other people. Racewalking came into the conversation because it is a meaningful part of my life. I would like to connect with that community. Employment and study are also goals I want to work towards.

These goals give the communication work a purpose beyond answering an agency. I want to ask for help more comfortably, have a conversation without worrying afterwards that I crossed a line, and build relationships around things I enjoy.

The proposed fortnightly frequency is a discussion to develop through the report and planning process. The session arrangements, commencement, duration and funding still need confirmation.

Isabel spoke about assembling the report the following Monday, which would be 16 November, with the next day also mentioned as a possibility and 23 November given as a later outer date. The follow-up email asks her to confirm the expected timing.

Making the evidence useful for NDIS consideration

I want the NDIS to be able to understand what a support helps me do. My experience provides examples. Clinical assessment can examine those needs, the approaches worth considering and the support required to use them effectively.

NDIS guidance on assistive-technology evidence asks providers to connect a participant's needs and goals with the proposed technology, explain options considered or trialled, address value and identify risks. Its general assessment guidance also asks how technology works with the person's other supports. That gives us practical questions to work through.

NDIS guidance on preparing assistive-technology evidence · NDIS guidance on general assistive-technology assessments

I would like to explore outcomes such as preparing a clearer appointment agenda, understanding the action requested in a form, retaining the next steps after a call and sending one coherent follow-up with the relevant evidence. These are proposed measures for discussion with Isabel, rather than results already demonstrated in a formal trial.

My earlier occupational-therapy work with Michael Lewis and my planned Vision Australia assessment provide further context to coordinate. Their original reports should carry their own findings. Isabel's report can contribute her discipline's assessment of communication and participation.

Thank you, Isabel

I spend a great deal of time documenting difficult encounters. I also want to record an encounter in which I felt listened to and could see possibilities for further work.

Isabel repeated questions, explained tasks, recognised strategies and brought us back to the assessment when my thoughts went elsewhere. She explored the practical difficulties alongside my interests and goals. She was willing to seek advice where the technology discussion needed it.

I came away wanting to continue speech pathology. I wanted to send her a useful account of the conversation, follow through on the report and turn the discussion into work we can do together.

To Isabel and the practice team who helped arrange the appointment: thank you. Having someone take the time to understand how I communicate makes a difference.

My words are still mine. With appropriate support, I can make them easier to organise, easier to hear and easier to act on.


Sources and related reading

  • Participant source pack, 9 November 2026. Assessment source excerpts and notes sent to Isabel. Relevant excerpts include 2–4, 39–46, 63–68, 76, 79–88. Quotations are from the supplied text, with the transcription and attribution limitations described above.
  • Sent follow-up, 9 November 2026, 12:56 pm AEDT. “Thank you — 9 November speech pathology assessment notes and follow-up,” with both named attachments. Sending verified against the outgoing Gmail record.
  • Communicate Speech Pathology's team page, including Isabel's professional profile.
  • Beyond the Inbox: How Technology Restores My Voice, the earlier account that helped me prepare for this conversation.

Popular Posts